AI-derived value distribution system, data contribution verification system, method, and program

Abstract

We provide a social infrastructure system to fairly distribute the wealth generated by AI to humanity while protecting privacy. [Solution] The data contribution verification system 100 includes a data preprocessing unit 10 that performs differential privacy processing on raw data received from a user terminal 200, adding noise based on a predetermined privacy budget to generate an anonymized feature vector that reduces personal identifiability while maintaining statistical characteristics; a data value calculation unit 20 that uses a machine learning model to calculate the degree of influence that the anonymized feature vector has on improving the accuracy of the learning model and outputs it as a contribution score; a proof generation unit 30 that generates a zero-knowledge proof indicating that the contribution score was calculated based on a valid algorithm; and a ledger linkage unit 40 that transmits the contribution score and the zero-knowledge proof as a transaction to the blockchain network 500.

Description

Technical Field

[0001] This disclosure relates to a technology for distributing value derived from AI (Artificial Intelligence).

Background Art

[0002] Due to the rapid development of AI and robotics, while personal data has value as a learning resource for AI, the high degree of automation in the production process is transforming the conventional labor market. However, the current social and economic system is designed on the premise of the value of human labor, and the technical foundation for properly evaluating the legitimate contribution of personal data and for properly and autonomously redistributing the wealth generated by automation to society as a whole has not yet been established.

Disclosure of the Invention

Problems to be Solved by the Invention

[0003] This disclosure provides a system, method, and program that prevent the concentration of wealth due to AI and automation, and fairly and autonomously distribute wealth to data providers and society as a whole based on value other than labor.

Means for Solving the Problems

[0004] One aspect of this disclosure is a system that verifiably records the contribution degree of data in order to realize a fair distribution of the value created by AI based on the data provided from a plurality of user terminals, configured as a computer including a processor and a memory, a data preprocessing unit that executes a privacy protection process on the data received from the user terminal to generate anonymized features, a data value calculation unit that calculates the influence degree of the anonymized features on improving the accuracy of the learning model using a machine learning model and outputs the calculation result as a contribution score, a proof generation unit that generates proof information proving that the calculation of the contribution score is performed based on a legitimate algorithm, The data contribution verification system is characterized by comprising a ledger linkage unit that transmits the contribution score and the proof information to a distributed ledger on which the token issuance process is performed.

[0005] Another aspect of this disclosure is, A method for recording the contribution of data in a verifiable manner, in order to achieve a fair distribution of value created by AI based on data provided from multiple user terminals, A computer equipped with a processor and memory, A data preprocessing step involves performing privacy protection processing on the data received from the user terminal to generate anonymized features, A data value calculation step that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation step that generates proof information to prove that the calculation of the contribution score was performed based on a valid algorithm, The data contribution verification method is characterized by performing a ledger linkage step of transmitting the contribution score and the proof information to a distributed ledger on which the token issuance process is executed.

[0006] Another aspect of this disclosure is, A program to make a computer function as a data contribution verification system that records the contribution of data in a verifiable manner in order to achieve a fair distribution of wealth (value) created by AI based on data provided from multiple user terminals, The aforementioned computer, A data preprocessor performs privacy protection processing on the data received from the user terminal to generate anonymized features. A data value calculation unit that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score. A proof generation unit that generates proof information to prove that the aforementioned contribution score was calculated based on a valid algorithm, and This is a data contribution verification program that functions as a ledger linkage unit, which transmits the aforementioned contribution score and the aforementioned proof information to a distributed ledger where the token issuance process is performed.

[0007] Another aspect of this disclosure is, A system that uses smart contracts deployed on a blockchain network to distribute value derived from production activities by artificial intelligence or robots to multiple users, A macroeconomic interface that communicates with external national AI funds, tax systems, or central bank digital currency systems to detect and tokenize the growth in GDP or profits of AI-related companies resulting from the spread of AI, and supplies them as liquidity to the basic income pool. A purchasing power parity-adjusted oracle that monitors market prices in real time for a basket of essential goods, including energy, food, and housing costs, and obtains a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions, This AI-derived value distribution system is characterized by comprising: a dividend logic that uses tokens accumulated in the basic income fund pool as its source, dynamically calculates the amount of benefits per person necessary to maintain real purchasing power based on the real cost of living index obtained from the purchasing power parity adjusted oracle, and automatically distributes it to the user's wallet address.

[0008] Another aspect of this disclosure is, A method of distributing value derived from production activities by artificial intelligence or robots to multiple users using smart contracts deployed on a blockchain network, A macroeconomic linkage step involves a computer equipped with a processor and memory communicating with an external national AI fund, tax system, or central bank digital currency system to detect and tokenize the growth in GDP or profits of AI-related companies resulting from the proliferation of AI, and supplying this tokenized liquidity to a basic income pool. A purchasing power parity adjustment step that monitors market prices in real time for a basket of essential goods including energy, food, and housing costs, and obtains a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions, The AI-derived value distribution method is characterized by including a dividend execution step of dynamically calculating the amount of benefits per person necessary to maintain real purchasing power based on the acquired real cost of living index, using tokens accumulated in the basic income fund pool as the source, and automatically distributing it to the user's wallet address.

[0009] Another aspect of this disclosure is, A program that enables a computer to function as an AI-derived value distribution system that uses smart contracts deployed on a blockchain network to distribute value derived from production activities by artificial intelligence or robots to multiple users, To the aforementioned computer, A macroeconomic linkage function that communicates with external national AI funds, tax systems, or central bank digital currency systems to detect and tokenize the growth in GDP or profits of AI-related companies resulting from the spread of AI, and supplies them as liquidity to the basic income pool. A purchasing power parity adjustment function that monitors market prices in real time for a basket of essential goods, including energy, food, and housing costs, and obtains a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions, and This is an AI-derived value distribution program that implements a dividend logic function that uses tokens accumulated in the aforementioned basic income fund pool as its source, dynamically calculates the amount of benefit per person necessary to maintain real purchasing power based on the acquired real cost of living index, and automatically distributes it to the user's wallet address. [Effects of the Invention]

[0010] According to the present disclosure, it is possible to prevent the concentration of wealth due to AI and automation, and to fairly and autonomously distribute wealth to data providers and society as a whole based on values other than labor.

Brief Description of Drawings

[0011] [Figure 1] It is a block diagram showing the overall configuration of the data contribution degree verification system according to the first embodiment. [Figure 2] It is a diagram showing the internal processing flow in the data value calculation unit. [Figure 3] It is a diagram showing the configuration of the token supply control logic and the inflation / burn mechanism. [Figure 4] It is a diagram showing the on-chain and off-chain cooperation process. [Figure 5] It is a diagram showing the audit trail generation process and the Merkle tree structure. [Figure 6] It is a diagram showing the cooperation between the privacy setting screen and the preprocessing pipeline in the user terminal. [Figure 7] It is a diagram showing the distribution control flow by the dividend scheduler and the fraud detection module. [Figure 8] It is a diagram showing the KYC / AML check and the cashing flow in cooperation with the settlement gateway. [Figure 9] It is a sequence diagram of the generation and verification of zero-knowledge proof. [Figure 10] It is a diagram showing the cooperation with the government identification infrastructure and the automatic transfer flow. [Figure 11] It is a diagram showing the parameter update flow by governance voting. [Figure 12] It is a configuration diagram of the log reception and normalization process in the API gateway. [Figure 13] It is a diagram showing the configuration and fund flow of the basic income distribution system. [Figure 14] It is a detailed configuration diagram of the consent management system. [Figure 15]This diagram shows the detailed flow of interconnection with heterogeneous AI providers and log normalization. [Figure 16] This figure shows the overall configuration and value cycle flow of the universal high-income realization system according to the second embodiment. [Figure 17] This is a flowchart showing the processing procedure for the data contribution verification method according to the third embodiment. [Figure 18] This is a flowchart illustrating the processing procedure for the universal high-income realization method according to the fourth embodiment. [Modes for carrying out the invention]

[0012] The present disclosure will be described in detail below with reference to the attached drawings. (Regarding terminology) Wealth (Value) Distribution: This is the process of automatically and fairly returning the economic benefits (GDP growth and corporate profits) created by AI's production activities to users through smart contracts as rewards (merit-based) and / or dividends that guarantee a basic standard of living (universal-based) according to the data's contribution. In this disclosure, "wealth" and "value" may be interpreted interchangeably.

[0013] Proof Information: This data mathematically and cryptographically guarantees that there has been no fraud or tampering in the calculation of the contribution score. This information allows for verification of the validity of the score calculation. A distributed ledger, such as a blockchain, is a system that does not rely on a single managing entity, but instead shares data among network participants and records and manages it with tamper resistance. In the system disclosed here, it is mainly used for issuing and recording tokens and verifying proof information. Differential privacy is a mathematical definition that prevents it from being determined from the output whether a dataset contains data on a specific individual. This is usually achieved by adding noise that follows a Laplace or Gaussian distribution. Privacy budget (ε): In differential privacy, this parameter controls the level of privacy protection. A smaller value results in greater noise and stronger privacy protection, but at the expense of data usefulness (accuracy). Anonymized feature vector: This is a sequence of numerical values ​​(features) extracted from raw data (images, text, etc.) that has been processed to reduce its ability to identify individuals. The AI ​​model processes this vector as input. Zero-knowledge proof (ZKP): A cryptographic technique that proves a proposition is true without revealing any of the secret information (raw data or model parameters) that underlies it. Smart contract: A contract (program) that is automatically executed on the blockchain when pre-programmed conditions are met. A Merkle tree is a tree structure that hierarchically combines data summaries (hash values). The hashes of the data at the lowest level (leaf nodes) are paired and hashed, and this process is repeated to ultimately generate a single hash value (root hash). Root Hash: This is the hash value at the top of a Merkle tree. A single Root Hash can represent the integrity of all data contained within the tree. A Merkle proof is the minimum set of hash values ​​required to prove that a particular piece of data is included in a Merkle tree. Using this proof, it is possible to prove the existence of specific data without knowing the entire tree.

[0014] Shapley Value: A concept in cooperative game theory, it is a calculation method for fairly distributing the contribution of each participant (each data point) to a given outcome (improvement in the accuracy of an AI model). Gradient-based influence analysis (GRA) is a method that estimates how much the training data influenced the model's parameter determination by analyzing the gradient (rate of change) of the training data relative to the loss function of the trained model. It tends to have lower computational costs than Shapley values.

[0015] Dynamic Adjustment: This refers to a mechanism that automatically changes parameters according to the situation. In this context, it refers to a function that increases or decreases the level of privacy protection (ε) in real time according to the nature of the data and the required accuracy.

[0016] Inflation control logic: This is an algorithm that automatically limits or adjusts the issuance pace to prevent the token's value from plummeting due to an excessive increase in the supply. Burning: This operation permanently removes issued tokens from the market by sending them to an address where no one possesses the private key (an invalid address). This reduces the supply and increases the value of the tokens. Price feed: This is a mechanism that obtains current token price information from external cryptocurrency exchanges, etc., and incorporates it into a smart contract.

[0017] Off-chain refers to processing performed outside the blockchain network. In this context, it refers to computationally intensive tasks such as log analysis and integrity checks.

[0018] Oracle: A general term for middleware and services that accurately transmit real-world data (such as verification results and price information) located outside the blockchain (off-chain) to smart contracts within the blockchain (on-chain). Simulation function: This function predicts and displays, in graphs and numerical values, how much token dividends can be expected in the future if data is provided with the current settings (noise level, etc.).

[0019] Scarcity: This refers to the degree to which there are few providers with that data type. The higher the scarcity, the higher the score. Recency: This refers to the freshness of the data. In AI learning, the latest data is highly valuable and therefore becomes a weighting factor.

[0020] Ground Truth: This is historical data of past dividend payouts and the resulting improvements in AI accuracy, which is used as the "correct" answer when a machine learning model learns its weights. Statistical anomaly detection: This is a method for statistically detecting behavior that deviates from normal data patterns (e.g., a particular user having an exceptionally high score).

[0021] Dividend Suspend: This process involves temporarily suspending asset transfers for transactions suspected of being fraudulent, without confirming the transaction, until human verification or additional automated verification is completed.

[0022] Explainability (XAI): This refers to the ability and technology to present the rationale behind AI's decision-making processes and results in a way that humans can understand. This makes it possible to demonstrate that the AI's assessments are based on evidence, rather than being a black box, and contributes to increased user trust in the system. Specific methods used include technologies such as SHAP and LIME.

[0023] Dashboard: This is a user-facing management screen that displays a breakdown and history of contribution scores in graphs and other formats. Governance function: This is a voting and proposal mechanism for determining the operating rules of an organization or system. On a blockchain, it is common for voting rights to be exercised in proportion to the amount of tokens held.

[0024] API Gateway: A server that acts as a "gatekeeper," receiving communication requests from external systems in a centralized manner, performing authentication and data transformation, and then passing them on to internal systems. Normalization adapter: A program component that converts data formats that differ from provider to provider (such as JSON structure) into a standard format within the system. Government ID Linkage Interface: This is a functional component that connects the public personal authentication service with this system using the My Number Portal API, etc.

[0025] KYC (Know Your Customer): This refers to the identity verification process that is mandatory when using financial services. Here, this process is automated and made more stringent through government ID linkage. Funding pool: This is a savings address on a smart contract where a portion of the usage fees for AI services and revenue from data trading are accumulated for dividends. Basic Income Distribution Logic: This is a programmatic process that distributes tokens regularly and evenly to registered unique users (duplicates have been eliminated using government IDs, etc.), regardless of their level of contribution.

[0026] Consent Management User Interface: This is an operation screen that allows users to visually adjust the accuracy and scope of the data provided, not only by turning categories ON / OFF, but also by using sliders and other controls. Granularity refers to the level of detail and precision of data. Examples include the difference between "GPS coordinates (high granularity)" and "city name (low granularity)" in location information, or between "product name (high granularity)" and "genre (low granularity)" in purchase history.

[0027] Filtering logic: This is a programmatic process that dynamically applies necessary masking or deletion processes to the input raw data based on the user's consent settings. Forget Request: Triggered by a user's withdrawal of consent, this is a signal that instructs various modules within the system (such as the learning database and cache) to delete or discontinue the use of data related to that user.

[0028] Standard log format: This is a data description format commonly used within the system. It eliminates inconsistencies in notation across different providers and maintains user IDs, data types, usage amounts, etc., in a unified structure. Normalization adapter group: A collection of conversion programs (mappers) that convert an external format (input) into an internal standard format (output).

[0029] Replay Attack: This attack method involves eavesdropping on and saving legitimate communication data (in this case, AI usage logs), and then retransmitting it later, causing the system to mistakenly believe it is legitimate communication. Window check: This process verifies whether the timestamps included in the log are within a certain acceptable range (window) from the current time. Older logs outside this range are discarded, thus serving as a countermeasure against replay attacks.

[0030] Macroeconomic Interconnection Interface: This is a gateway that connects to external national budget and financial systems via APIs, and tokenizes and imports funds. Purchasing Power Parity (PPP) Adjustment Oracle: A system that accurately transmits real-world price information within the blockchain. Basket of essential goods: This represents the average price of essential goods for daily life (food, energy, housing, etc.).

[0031] Reputation tokens (Soulbound Tokens / SBTs): These are tokens that cannot be transferred or bought / sold to others. They function as proof of a person's accumulated credibility and achievements. Immutable: Once written, it cannot be changed or deleted by anyone (not even the developer). Censorship resistance refers to the property of a system that prevents central administrators or specific individuals in power from arbitrarily excluding or manipulating certain transactions or participants.

[0032] Universal High Income (UHI): Unlike traditional basic income (minimum living guarantee), UHI is a concept that uses surplus profits generated by productivity improvements through AI and robots to distribute an amount sufficient to live a decent and fulfilling life (high income).

[0033] Artificial General Intelligence (AGI): Artificial intelligence that is not limited to specific tasks and possesses a wide range of cognitive abilities equal to or greater than those of humans. Artificial Superintelligence (ASI): An artificial intelligence that surpasses the combined intelligence of all humanity and is capable of autonomously making scientific discoveries and optimizing production activities.

[0034] Real Cost of Living Index: In a society where production costs decrease due to automation by AI, this index shows fluctuations in the actual costs (energy, food, housing, etc.) necessary for humans to maintain a certain standard of living, rather than nominal prices. Social capital is a concept that refers not to monetary assets, but to an individual's social credibility, reputation, and the richness of their relationships within a community. In the system disclosed here, it is quantified as the amount of SBT (Reputation Tokens) accumulated.

[0035] National AI Fund: A sovereign wealth fund established by a government or public institution, which holds and manages assets such as the economic growth resulting from automation by AI and robots (GDP increase, AI tax, or corporate excess profits). In this disclosure, it refers to an external source of financial resources that connects through a macroeconomic interface and supplies its investment gains and dividends as liquidity to the basic income pool by tokenizing them.

[0036] Robot Tax Collection System: A public tax infrastructure system that automatically calculates and collects taxes on excess profits gained by companies through the operation of robots and AI that replace human labor, or on the amount equivalent to the reduced labor costs. In this disclosure, it refers to an external system that connects through a macroeconomic interface and transfers the collected tax revenue to a basic income fund pool as digital currency or tokens.

[0037] A central bank digital currency (CBDC) issuance system is a payment infrastructure system issued by a nation's central bank that manages and circulates digital currencies (such as digital yen) that have value as legal tender. In this disclosure, it refers to an external financial infrastructure that, through a macroeconomic linkage interface, immediately and automatically transfers collected AI taxes and national budgets to a basic income fund pool as programmable digital currency.

[0038] 1. First Embodiment [Technical field] This disclosure relates to data processing technology in a decentralized network environment, and in particular to a technology that quantitatively calculates the contribution of data to a machine learning model while cryptographically protecting individual privacy, and makes the validity of the calculation results verifiable on the blockchain. [Background technology] In recent years, with the development of AI technologies, including deep learning, the demand for high-quality training data has surged. The value of personal data generated not only from data held by companies and organizations, but also from individual devices (smartphones, IoT devices, etc.) is increasing. Traditionally, reward systems for data providers have typically included fixed rewards based on the amount of data provided or points awarded based on simple access logs. However, since the contribution to improving the accuracy of AI models depends not on the quantity of data but on its quality (scarcity, diversity, and accuracy), traditional systems have made it difficult to achieve a fair distribution of value. Furthermore, accurately measuring the contribution of data requires analyzing raw data on the server side, which carries the risk of privacy violations. On the other hand, excessive processing and anonymization of data for privacy protection impairs its practical usability (effectiveness for AI learning), creating a trade-off. In addition, contribution scores calculated by centralized servers tend to be black boxes, and there was a transparency issue where data providers could not verify their legitimacy. [Purpose of this disclosure] This disclosure has been made in view of the above-mentioned issues, and its purpose is to provide a data processing platform that simultaneously achieves the following three points. (1) Balancing privacy and practicality: Using technologies such as differential privacy, the data is converted into a format that allows for the calculation of contributions to the AI ​​model while keeping individuals unidentifiable. (2) Proof of the validity of the calculation: Prove cryptographically that the contribution score was calculated using the correct algorithm without disclosing the raw data (zero-knowledge proof). (3) Ensuring scalability and transparency: Enabling the recording and auditing of a large number of transactions in an immutable manner without straining the blockchain's processing capacity.

[0039] A. Overall system configuration Figure 1 shows the overall configuration of the data contribution verification system (hereinafter also referred to as "the System") 100 according to this embodiment. The System 100 consists of multiple user terminals 200 connected via a network NW, an external AI service provider 300, and components on a blockchain network 500.

[0040] The core computing nodes of this system 100 comprise the following functional blocks: Data preprocessing unit 10: It applies differential privacy processing to the raw data received from the user terminal 200 to generate an anonymized feature vector. Data Value Calculation Unit 20: Calculates the contribution score of the data using a machine learning model. Proof generation unit 30: Generates a zero-knowledge proof (ZKP) that proves the validity of the contribution score. Ledger linkage unit 40: Sends the generated score and proof as a transaction to the blockchain network 500. Smart contracts 400 are deployed on blockchain network 500, and token issuance and dividend processing are executed conditional on the successful verification of ZKP. It also interacts with API gateway 60 (see Figure 12) and off-chain verification node 70 (see Figure 4).

[0041] Each functional block constituting this system 100 (data preprocessing unit 10, data value calculation unit 20, proof generation unit 30, ledger linkage unit 40, etc.) is a specific means of realization through the cooperation of hardware resources and software. Physically, each of these units is constructed as one or more computer systems equipped with a CPU (Central Processing Unit), main memory (RAM), ROM, auxiliary storage devices (SSD, HDD, etc.), communication interfaces, and bus lines connecting them.

[0042] The functions of each section are realized when various programs, such as data contribution verification programs stored in auxiliary storage devices, are read into main memory and interpreted and executed by a processor such as a CPU. In particular, in the implementation of the data value calculation unit 20, which requires a large amount of matrix calculations, and the proof generation unit 30, which requires advanced cryptographic calculations (such as elliptic curve calculations), a configuration may be adopted that uses a general-purpose CPU in combination with hardware accelerators such as a GPU (Graphics Processing Unit), TPU (Tensor Processing Unit), or FPGA (Field-Programmable Gate Array) to parallelize and speed up processing.

[0043] Furthermore, from a privacy perspective, it is desirable that a portion of the data preprocessing unit 10 or the certificate generation unit 30, which directly handles raw data, be executed on a processor compatible with TEE (Trusted Execution Environment) (for example, Intel SGX or AMD SEV). By using TEE, data in memory is encrypted and isolated at the hardware level, guaranteeing a secure computing environment in which even OS administrators or cloud service providers cannot pry into the raw data being processed.

[0044] B. Detailed process of data valuation Referring to Figure 2, the processing of the data value calculation unit 20 will be explained. The data value calculation unit 20 takes the anonymized feature vector output from the data preprocessing unit 10 as input and performs the following steps. First, the feature extractor 21 extracts high-dimensional features from the input data. Next, the weighting model 22 calculates weight coefficients based on the indicators of data scarcity, recency, quality, and usage context. Finally, the score calculation engine 23 calculates the marginal contribution that the data has to improving the accuracy of the AI ​​model using the Shapley Value or gradient-based influence analysis and outputs a contribution score S. In addition, the explainability module 24 analyzes which features made a positive contribution to the calculated score S, generates explanatory information (XAI data), and sends it to the user dashboard.

[0045] C. Token Economics and Inflation Control Figure 3 shows the token supply control system implemented in the smart contract 400 or ledger linkage unit 40. The inflation control logic 410 takes the current total token supply, demand indicators, and price feeds obtained via an external oracle as inputs to determine the mint rate for new tokens. If the market supply is determined to be excessive, the burn mechanism 420 is activated, either using a portion of the system revenue to buy back the tokens or sending undistributed tokens to a burn address for destruction. In addition, a portion of the system revenue is accumulated in a basic income fund pool (not shown), and the basic income distribution logic distributes tokens for basic living benefits to all registered users who meet the conditions on a regular and equal basis.

[0046] D. On-chain and off-chain collaboration Figure 4 illustrates a collaborative structure that balances scalability and security. Large volumes of usage logs from the AI ​​service provider 300 are not directly recorded on the blockchain 500, but are analyzed by the off-chain verification node 70. The verification node 70 verifies the integrity and confidentiality of the logs and transmits the results to the smart contract 400 via the oracle 510. The smart contract 400 sets the "approval signal" from the oracle 510 as a mandatory condition (trigger) for token distribution, thereby preventing distribution based on fraudulent logs.

[0047] E. Audit trails and Merkle trees As shown in Figure 5, the ledger linkage unit 40 constructs a Merkle tree 520 using the hash values ​​of numerous transactions (contribution score and proof) that occurred within a certain period as leaf nodes. Only the root hash 521, which is the root of this tree, is recorded in the blockchain 500. This saves on-chain capacity. Upon request from an external auditing body 950, the ledger linkage unit 40 generates and outputs a Merkle proof 522 that mathematically proves that a particular transaction is included in the root hash.

[0048] F. Privacy Settings and Simulation Figure 6 shows the privacy settings UI 210 displayed on the user terminal 200. Users can adjust the differential privacy parameter (privacy budget ε) by operating sliders, etc. The simulation function 211 on the UI 210 calculates the amount of noise based on the selected ε value and displays the degree of data degradation and the predicted increase or decrease in token dividend amount in real time on a graph. The data preprocessing unit 10 receives this setting value and controls the noise addition process by dynamic adjustment logic 11. It also has a feedback loop that automatically corrects the ε value if the decrease in the accuracy of the contribution score exceeds a threshold.

[0049] G. Dividend control and fraud detection In Figure 7, the smart contract 400 includes a dividend scheduler 430 that controls the timing of dividend distributions (real-time, daily, monthly, etc.). Prior to the dividend execution process (transaction creation), a fraud detection module 440 is placed. This module uses a statistical anomaly detection algorithm to monitor sudden spikes in scores and bot-like access patterns. If an anomaly is detected, the dividend processing immediately enters a suspended status and branches to a re-verification flow by an administrator or additional verification program.

[0050] H. Payment and Government ID Linkage Figures 8 and 10 illustrate the connection to real-world financial and administrative systems. In Figure 8, when a request for token exchange for fiat currency arises, the settlement gateway 800 is activated, performs KYC (Know Your Customer) and AML (Anti-Money Laundering) checks, and then executes the bank transfer. As shown in Figure 10, this system is equipped with a government ID linkage interface 600. Interface 600 connects via API to the government identification infrastructure 610, such as My Number, and links the user's wallet address and government ID while maintaining confidentiality using a hash function or ZKP. This makes it possible to automatically convert distributed tokens into cash and transfer them to a public funds receiving account.

[0051] I. Generation and Verification of Zero-Knowledge Proofs Figure 9 shows the interaction between the proof generation unit 30 and the smart contract 400. The proof generation unit 30 (Prover) takes secret information (Witness: raw data, model parameters) and public information (PublicInput: score, model hash) as input, executes the proof circuit 31 (Arithmetic Circuit), and generates proof data π. The smart contract 400 (Verifier) ​​executes the verification function Verify(Proof, Public Input). Only if this function returns True is the data value calculation deemed valid, and the process proceeds to the token issuance logic.

[0052] Furthermore, in order to reduce computational costs, the proof generation unit 30 uses an efficient proof scheme such as zk-SNARKs (Zero-Knowledge Succinct Non-Interactive Argument of Knowledge) to generate zero-knowledge proofs. Alternatively, instead of proving all computational steps of the AI ​​model, a lightweight circuit may be adopted that only proves the consistency between the hash value of the input data and the output score, and the validity of the model's fingerprint. Alternatively, the Optimistic Rollup technique may be applied to omit proofs under normal circumstances and generate and verify detailed proofs (fraud proofs) only when suspicion of fraud arises.

[0053] J. Governance and API Gateway In Figure 11, the governance function 450 accepts votes from token holders. Based on the voting results, the inflation rate and weighting model parameters are automatically updated. In Figure 12, the API gateway 60 receives logs in different formats from multiple external AI providers 300A, 300B, etc. The normalization adapter 61 unifies these into a standard format, and the signature verification unit 62 verifies the digital signature and timestamp before passing the data to the next stage of processing.

[0054] K. Details of the detailed consent management system Figure 14 shows the detailed configuration of the Consent Management System. In this system, users, who are data providers, can intuitively and precisely control their privacy settings through the consent management UI 220 implemented on the user terminal 200.

[0055] The consent management UI220 displays toggle switches to enable or disable data provision for each data category, such as "healthcare," "location information," and "finance / purchasing," as well as sliders (granularity settings) to adjust the level of detail (granularity) of the data provided. For example, a user can select "provide" location information but flexibly specify that the accuracy be limited to "prefecture level (low granularity)" rather than "GPS coordinate level (high accuracy)."

[0056] User-defined consent information is sent to the "Consent History DB 16" within the data preprocessing unit 10 and recorded in an tamper-proof format along with a timestamp. The "Filtering Logic 15," also located within the data preprocessing unit 10, performs gateway processing on incoming raw data by referencing the consent history DB 16 in real time. Specifically, data in categories that the user has not consented to (OFF setting) is physically blocked or nulled before being passed to subsequent processing, thereby reliably preventing unintended data leakage.

[0057] Furthermore, when a user operates the "Revoke Consent" button on the consent management UI220, the status in the consent history DB16 is updated, and at the same time, a "forget request" is sent from the filtering logic 15 to the data value calculation unit 20, etc. As a result, the system stops using the user's past data and executes a process to exclude it from future machine learning model update cycles.

[0058] Furthermore, this consent setting is also linked to the token dividend amount. The data value calculation unit 20 calculates the contribution based on the range and granularity of the "permitted data" that has passed the filtering logic 15. Therefore, users who provide all categories with high accuracy are given a high weighting coefficient, while users who block some categories or lower the granularity are given a coefficient corresponding to those restrictions.

[0059] L. Interconnection with heterogeneous AI providers and log normalization For this system 100 to function as a common platform for the entire AI ecosystem, it is necessary to accurately and efficiently collect usage logs from different external AI service providers 300 (300A, 300B, etc.) such as Google, OpenAI, and Microsoft. To solve this problem, an advanced API gateway 60 (see Figure 12 or Figure 15) is placed at the boundary between the system and the external providers.

[0060] When API Gateway 60 receives usage logs (including metadata of inference requests, model IDs, hashes of used features, etc.) sent from each provider, it first performs security verification. Specifically, it performs digital signature verification (ECDSA, etc.) using a public key on logs signed with private keys distributed to each provider in advance to confirm the authenticity of the sender. At the same time, it verifies the timestamps contained in the logs and performs a window check to prevent replay attacks (attacks that artificially inflate dividends by resending past logs).

[0061] Logs that have passed security verification often have different data formats (JSON schema, XML, CSV, etc.) depending on the provider. Therefore, API Gateway 60 implements a "normalization adapter group" to absorb the differences between providers. The normalization adapter analyzes each company's proprietary format and converts (maps) it to the "Standardized Log Format" defined within this system. This standard format includes a unique user identifier (hashed), the data category used, a provisional value for contribution, and a transaction ID.

[0062] Only the normalized log data is passed on to the subsequent data value calculation unit 20 and smart contract 400. This allows the system's internal logic to perform contribution calculation and dividend execution using a unified algorithm, without being affected by changes in the specifications of external providers. This architecture can also be adapted in the future when new AI service providers enter the market, simply by adding the corresponding new adapters, dramatically increasing the system's scalability and availability.

[0063] M. Basic Income Dividend System Figure 13 shows the structure of the basic income (basic living benefit) distribution system. This system 100 aims for sustainable data collection and community expansion by using not only performance-based rewards (merit-based) according to data contribution, but also a basic income (universal-based) mechanism that evaluates participation in the ecosystem itself.

[0064] Within smart contract 400, there is a basic income fund pool 460 that automatically accumulates a portion of the system's revenue (for example, a percentage of usage fees and transaction fees from AI service providers 300). The inflow of funds into this pool is enforced by the program and cannot be arbitrarily stopped by any particular administrator.

[0065] The basic income distribution logic 470 implemented in the smart contract 400 checks the token balance accumulated in the principal pool 460 when a predetermined distribution cycle (e.g., monthly) arrives, divides this by the number of eligible registered users to calculate the equal distribution amount per person. Then, regardless of the level of each user's data contribution score, it automatically distributes the same amount of tokens to all eligible users.

[0066] Here, it is desirable that the dividend logic 470 not only "registers," but also determines whether the user has met certain minimum data provision conditions (e.g., sending logs at least once a month and synchronizing healthcare data) and whether their existence has been verified through the government ID linkage interface 600 (see Figure 10). This will prevent fraudulent receipt of benefits by bots (Sybil attacks), encourage participation from a wide range of general users, and facilitate the collection of diverse long-tail data essential for improving the generalization performance of the AI ​​model.

[0067] According to the first embodiment, by utilizing cryptographic technologies such as differential privacy and zero-knowledge proofs, it becomes possible to fairly evaluate the qualitative contribution to AI models while simultaneously protecting the privacy of data providers and ensuring transparency in the evaluation process. In addition, by combining an autonomous token distribution mechanism using smart contract 400 with a basic income distribution logic 470 that applies to all participants, it is possible to prevent the monopolization of wealth by a few administrators or corporations and realize a decentralized economic system that fairly and sustainably distributes the value created by AI and automation to data providers and society as a whole based on new value standards other than labor.

[0068] 2. Second Embodiment [Technical field] This disclosure relates to a social infrastructure system utilizing blockchain and smart contracts, and in particular to a decentralized economic system that realizes Universal High Income (UHI) by automatically distributing wealth (value) generated from production activities by general artificial intelligence (AGI) and robots, dynamically adjusts benefit amounts based on purchasing power parity, and visualizes non-monetary value (social capital) that does not depend on labor.

[0069] [Background technology] In recent years, the rapid advancements in artificial general intelligence (AGI) and artificial superintelligence (ASI), along with the proliferation of autonomous robots, have led to predictions of a future where a large portion of physical and intellectual labor is replaced by machines. This is expected to dramatically increase overall societal productivity, ushering in a zero-marginal-cost society where the marginal cost of goods and services is reduced to its absolute minimum, and ultimately leading to an abundance where necessities can be obtained without labor.

[0070] Traditionally, wealth redistribution and social security (pensions, welfare, etc.) have been designed primarily with taxation on human labor as the main source of funding. Furthermore, existing discussions on basic income (BI) have been limited to fixed payments that guarantee a minimum standard of living, and have not adequately considered the distribution of high incomes funded by the enormous wealth generated by AI (explosive GDP growth), nor mechanisms to maintain real purchasing power in response to rapid deflation (falling prices) caused by technological innovation.

[0071] Furthermore, if work ceases to be a human obligation, there is a challenge in how to quantify "purpose in life" or "social recognition" that replace wage labor, and how to build trust within communities—a lack of a foundation for non-monetary value exchange. In addition, there are concerns about the centralized risk that wealth creation by AI will be monopolized by a few large corporations and administrators.

[0072] [Purpose of this disclosure] A second embodiment of this disclosure has been made in view of the above-mentioned problems, and its purpose is to provide a socio-economic system that comprehensively realizes the following four points. (1) Automatic linkage with the macroeconomy: AI and robots will detect GDP growth and corporate profits, and in conjunction with external national funds and tax systems, that wealth will be automatically returned as liquidity to the basic income pool. (2) Maintaining real purchasing power: Monitor price fluctuations (deflation, etc.) caused by the spread of AI in real time and dynamically adjust the amount of benefits to guarantee a "real standard of living" rather than a nominal amount. (3) Visualization of social capital: Independent of monetary rewards, gratitude and recognition from others for volunteer work, creative activities, and care work will be recorded as "reputation tokens" to build a new social credit base. (4) Ensuring resistance to censorship: Decentralized autonomous organizations (DAOs) and immutable smart contracts will prevent the suspension of benefits or the monopolization of wealth by specific power holders, thereby guaranteeing a permanent safety net. This allows data providers to not only receive payment for their data, but also to enjoy the prosperity of the entire AI society as "shareholders," thereby gaining a sustainable and prosperous foundation for life after being freed from labor.

[0073] N. Universal High Income (UHI) Realization System (AI-derived Value Distribution System) While the first embodiment primarily described a mechanism for compensating individuals for providing data, the second embodiment will describe an expanded configuration as a social infrastructure that anticipates the explosive increase in productivity accompanying the spread of artificial general intelligence (AGI) and artificial superintelligence (ASI), and the eventual elimination of labor. The system of the second embodiment will have all the configurations of the system of the first embodiment.

[0074] Figure 16 shows the configuration of the UHI realization system according to the second embodiment. In this embodiment, the smart contract 400 works closely with external macroeconomic systems and non-monetary value exchange protocols to provide benefits for survival and evaluations for self-actualization in parallel.

[0075] The system of the second embodiment is equipped with a macroeconomic linkage interface 900. While the resource pool 460 in the first embodiment was primarily funded by data usage fees within the system, the interface 900 in the system of the second embodiment is API-connected to external national AI funds, robot tax collection systems, or central bank digital currency (CBDC) issuance systems. The growth in gross domestic product (GDP) resulting from automation by AI and robots, and the taxes on excess profits of AI companies, are automatically tokenized through this interface and continuously supplied as liquidity to the basic income resource pool 460. This makes it possible to provide benefits on a scale orders of magnitude larger (universal high income) compared to relying solely on data usage fees.

[0076] The macroeconomic linkage interface 900 features a standardized data reception protocol (e.g., an ISO20022-compliant financial message format or a RESTful API) that is independent of the implementation form of the external system. Specifically, it detects GDP growth and tax revenue by receiving signed transactions from the external system that include structured data (such as JSON) like "{Asset type: "AI_TAX_REVENUE", Amount: 10000000, Currency: "CBDC_JPY", Target period: "2030-Q1"}". This makes it possible to connect to systems that will be built in the future, by aligning them with the interface specifications of the system side in the second embodiment.

[0077] The aforementioned data structure example, "{Asset Type: "AI_TAX_REVENUE", Amount: 10000000, Currency: "CBDC_JPY", Period: "2030-Q1"}", illustrates a concrete example of standardized transaction information received by the macroeconomic linkage interface 900 from external national institutions and tax systems. Here, "AI_TAX_REVENUE" in "Asset Type" is a tag used to identify that the source of the funds is tax revenue or excess profits from automation by AI and robots, while "Amount" and "Currency" specify that the value transfer is in units of 10 million Japanese yen as central bank digital currency (CBDC). Furthermore, "2030-Q1" in "Period" specifies that the revenue was generated in the first quarter of 2030. Based on this information, the system accurately detects and tokenizes the national-scale economic growth and automatically supplies it as liquidity to the basic income fund pool.

[0078] Furthermore, to address the rapid deflation (fall in prices) resulting from the reduction in production costs due to AI, the system of the second embodiment includes a purchasing power parity (PPP) adjustment oracle 910. This oracle monitors the prices of a basket of necessities, such as energy prices, food prices, and housing costs, in real time and transmits them to the smart contract 400. The basic income dividend logic 470 does not simply distribute a fixed amount of tokens, but dynamically calculates the amount of benefit necessary to maintain or improve the real standard of living based on the price index obtained from the oracle. This makes it possible to maintain an equilibrium between the value of the currency and real purchasing power even in a hyperdeflationary phase.

[0079] The Purchasing Power Parity Adjusted Oracle 910 retrieves the food price (P_food), electricity price (P_energy), and communication cost (P_net) required for a standard daily calorie intake from APIs of major e-commerce sites and energy trading markets, and calculates a weighted average of these to determine the price of a basket of essential goods (P_basket). The Dividend Logic 470 determines the variable benefit amount (BI_real = BI_base × R) by multiplying the basic benefit amount (BI_base) by the current price ratio (R = P_basket / P_base) to the baseline price (P_base). This keeps the real purchasing power constant by decreasing the number of tokens during deflation and increasing them during inflation.

[0080] Furthermore, the system in the second embodiment implements the Proof of Contribution protocol 920 as a new evaluation axis distinct from monetary value exchange. In a society where work is no longer an obligation, activities that are difficult to evaluate using market principles, such as care work, creative activities, volunteering, and participation in local communities, are made visible through "gratitude" and "recognition" from others. Protocol 920 records peer-to-peer evaluations between users (sending "likes" and expressions of gratitude) and accumulates them in the social capital ledger 930 as non-transferable reputation tokens (Soulbound Tokens). This reputation score is managed independently of monetary rewards and functions as a measure of trustworthiness within the community and as a qualification for participation in specific projects.

[0081] These distribution mechanisms are managed by a censorship-resistant decentralized governance (DAO). Once deployed, the Smart Contract 400 code is immutable and fixed on the blockchain, preventing arbitrary shutdowns or freezes of benefits for specific individuals, even by specific government agencies or large tech companies. Any parameter changes are decided solely by decentralized voting by token holders and social capital holders, thus preventing centralized control and ensuring that it continues to function as a sustainable social security infrastructure.

[0082] According to the second embodiment, wealth generated by national-scale AI productivity improvements (GDP growth and robot tax revenue) is automatically reinvested into the basic income fund through the macroeconomic linkage interface 900, and the real benefit amount is adjusted in response to deflation and inflation using the purchasing power parity adjustment oracle 910, thereby realizing a sustainable "universal high income" that is not affected by the contraction of the labor market. In addition, the community contribution certification protocol 920 visualizes social contributions such as care work and creative activities, which are difficult to convert into monetary value, as inalienable reputation tokens, thereby building a new social credit base based on diverse values ​​other than labor, and enabling the fair and autonomous distribution of wealth in the AI ​​era to society as a whole.

[0083] 3. Third Embodiment: Data Contribution Verification Method In the first embodiment, the data contribution verification system 100 according to this disclosure was described from the perspective of each functional block (configuration element). In the third embodiment, a series of processing procedures executed by the system 100 using hardware resources (processor, memory, etc.) will be described from the perspective of a method invention with reference to the flowchart in Figure 17.

[0084] Figure 17 shows the overall flow of the data contribution verification method according to this embodiment. This process is initiated in parallel via the following two routes, triggered by data transmission from the user terminal 200 or log transmission from the external AI service provider 300.

[0085] <Step S1: Data reception and consent filtering process (user route)> When raw data is sent from the user terminal 200, the data preprocessing unit 10 receives it. The system refers to the consent history database 16 and determines whether valid consent has been obtained from the user regarding the category and granularity settings of the data. If the data contains information for which consent has not been obtained, the filtering logic 15 physically blocks or nullifies the data (S1).

[0086] <Step S2: Differential Privacy Application Process (User Root)> Next, the data preprocessing unit 10 performs differential privacy processing on the raw data that has passed through step S1, adding noise based on a predetermined privacy budget (ε). This generates an anonymized feature vector that reduces personal identification while maintaining the statistical characteristics (usefulness) of the data, and passes it on to the subsequent data value calculation unit 20 (S2).

[0087] <Step S10: Log reception and signature verification process (provider route)> On the other hand, when usage logs are sent from the external AI service provider 300, the API gateway 60 receives them. The API gateway 60 performs digital signature verification using a public key and a timestamp window check to confirm the authenticity of the sender and prevent replay attacks (S10).

[0088] <Step S11: Log Normalization Process (Provider Route)> If the verification is successful, the normalization adapters within the API gateway 60 convert (map) logs in different formats (JSON, XML, etc.) for each provider into the system's internal standard log format. The normalized data generated in this way is then passed to the subsequent data value calculation unit 20 (S11).

[0089] <Step S3: Data Value Calculation Process> Next, the data value calculation unit 20 takes the anonymized feature vector generated in step S2 or the normalized data generated in step S11 as input and calculates the influence of the data using a machine learning model. Specifically, it uses the Shapley Value or a gradient-based influence analysis algorithm to calculate a quantitative contribution score, taking into account weights such as the rarity and quality of the data (S3).

[0090] <Step S4: Zero-Knowledge Proof Generation Process> The proof generation unit 30 generates a zero-knowledge proof (ZKP) to prove that the contribution score calculated in step S3 was calculated based on the correct algorithm and input data. This process creates proof data (Proof) that mathematically guarantees only the correctness of the calculation while keeping the raw data and the internal parameters of the model secret (S4).

[0091] <Step S5: Transaction generation and on-chain verification process> The ledger linkage unit 40 generates a transaction containing contribution scores and proof data and sends it to the blockchain network 500. The smart contract 400 on the blockchain verifies the zero-knowledge proof in the received transaction by passing it through a verification function (Verifier). It also checks for an approval signal from the oracle as needed (S5).

[0092] <Step S6: Token Issuance and Dividend Distribution Process> If the verification in step S5 (and the "success" determination at the decision branch) is successful, the smart contract 400 executes the token issuance process (Mint). At this time, the inflation control logic 410 dynamically adjusts the issuance amount according to the market supply and demand situation. The issued tokens are sent to the user's wallet address, and a portion is automatically collected and accumulated in the basic income fund pool 460 (S6).

[0093] 4. Fourth Embodiment: UHI Realization Method (Value Distribution Method) In the second embodiment, the system configuration for realizing UHI was described. In the fourth embodiment, a series of processing procedures for providing benefits adapted to deflation (falling prices), using productivity improvements by AI as the source, will be described from the perspective of the method invention with reference to the flowchart in Figure 18. This process includes a loop process that is repeatedly executed at predetermined distribution cycles (e.g., monthly or daily) and an event process that is executed as needed in response to user actions.

[0094] <Step S21: Macroeconomic Linkage and Liquidity Supply Process> First, the system of the second embodiment (in other words, the smart contract 400 and the macroeconomic linkage interface 900) communicates with an external national AI fund, tax system, or central bank digital currency system to detect the economic growth (GDP growth rate or AI tax revenue) derived from production activities by AI and robots. The system of the second embodiment tokenizes the detected value and automatically injects and accumulates it as liquidity in the basic income pool 460 (S21).

[0095] <Step S22: Purchasing Power Parity (PPP) Acquisition Process> Next, the system of the second embodiment queries the Purchasing Power Parity (PPP) Adjusted Oracle 910 for current market price data in order to determine the dividend amount. Oracle 910 obtains real-time prices for a basket of essential goods consisting of energy, food, housing, communication costs, etc., and transmits the current price index (degree of inflation or deflation) to the smart contract 400 (S22).

[0096] <Step S23: Calculation Process for Actual Benefit Amount> Next, the basic income dividend logic 470 calculates the "real benefit amount" necessary to maintain a decent standard of living, based on the price index obtained in step S22. Here, if the cost of living decreases (deflation) due to the spread of AI, the benefit amount is dynamically determined so that real purchasing power is maintained or improved, while adjusting the nominal token amount (S23).

[0097] <Step S24: Universal Dividend Execution Process> The system of the second embodiment withdraws the tokens for the amount of the benefit determined in step S23 from the resource pool 460 and sends them simultaneously to the wallet addresses of all valid users whose existence has been verified via the government ID linkage interface 600. At this time, due to the irreversibility of the smart contract, benefit suspension or exclusion of individuals by specific authorities is not possible, and the distribution is carried out reliably as programmed (S24).

[0098] <Step S25: Social Capital Recording Process> In parallel with the monetary payments described in S21-S24 above, the system of the second embodiment keeps the community contribution verification protocol 920 running at all times. When users send thank-you messages or report the completion of volunteer activities, the system of the second embodiment verifies them and issues non-transferable reputation tokens (Soulbound Tokens) to record in the user's social capital ledger 930. This makes individual credit and contributions visible in a society where labor has disappeared (S25).

[0099] The method described herein can be implemented by running the program described herein (the Program) on a computer. The Program is stored on a computer-readable storage medium and executed when read from the storage medium. The Program may also be provided in the form of a non-volatile (non-transient) storage medium such as flash memory, or it may be provided via a network such as the Internet.

[0100] 5. Variations

[0101] In the first embodiment described above, an example was explained in which "differential privacy processing" is applied to raw data to generate an "anonymized feature vector" and then transmitted to a "blockchain network" along with a "zero-knowledge proof." However, the technical scope of this disclosure is not limited to this.

[0102] For example, the privacy protection processing performed by the data preprocessing unit 10 is not limited to differential privacy, but may also include k-anonymity, l-diversity, t-closeness, or concealment processing using secure multi-party computation or homomorphic encryption.

[0103] Similarly, anonymized features are not limited to vector (one-dimensional array) format, but encompass any form of features that a machine learning model can process, including scalar values, matrices, tensors, or graph-structured data.

[0104] Furthermore, the certification information generated by the certification generation unit 30 is not limited to zero-knowledge proofs (ZKP), but may be any authentication data that can guarantee the legitimacy of the calculation process, such as attestation by a trusted execution environment (TEE), digital signatures, or multi-signatures.

[0105] Furthermore, the distributed ledgers connected to the ledger linkage unit 40 are not limited to blockchain (chain-type structure), but refer to a broad range of distributed networks including DAG (Directed Acyclic Graph) type ledgers and consortium-type distributed ledger technology (DLT).

[0106] Although each embodiment of this disclosure has been described above with reference to the drawings, it goes without saying that this disclosure is not limited to the embodiments described above. It is clear to those skilled in the art that various modifications or alterations can be conceived within the scope of the claims, and these will naturally also fall within the technical scope of this disclosure. Furthermore, the elements of the embodiments described above may be combined in any way without departing from the spirit of the invention.

[0107] This specification discloses at least the following: (1) A system that records the contribution of data in a verifiable manner in order to achieve a fair distribution of value created by AI based on data provided from multiple user terminals, It is configured as a computer with a processor and memory, A data preprocessing unit performs privacy protection processing on data received from the user terminal to generate anonymized features, A data value calculation unit that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation unit that generates proof information to prove that the aforementioned contribution score was calculated based on a valid algorithm, A data contribution verification system characterized by comprising: a ledger linkage unit that transmits the contribution score and the proof information to a distributed ledger on which token issuance processing is performed.

[0108] (1) Action and effect This system receives data from user terminals and performs privacy protection processing (masking, anonymization, noise addition, etc.) on the data to generate anonymized features with reduced personal identifiability. These anonymized features are used as input to an AI model, ensuring data anonymity while maintaining the characteristics necessary for AI model training and evaluation. This reduces the risk of privacy violations and provides a foundation in which users can provide data with peace of mind. The data valuation unit takes the anonymized features as input, quantitatively calculates their impact on improving the accuracy of the machine learning model, and outputs the result as a contribution score. This score objectively evaluates the value based on the qualitative contribution to improving the performance of the AI ​​model, rather than the quantity of data provided, and serves as the basis for awarding data providers fair rewards (tokens) commensurate with their contributions. The proof generation unit generates proof information to mathematically and cryptographically guarantee that the calculated contribution score was performed based on a valid algorithm. This proof information and contribution score are recorded in a distributed ledger (a tamper-resistant ledger shared among network participants, such as a blockchain) via the ledger linkage unit. This prevents the score calculation process from becoming an opaque black box, allows the validity of the calculation to be verified without relying on a central system administrator or third-party institution, and ensures the reliability and transparency of the entire system. The distributed ledger is mainly used for token issuance and recording, and verification of proof information.

[0109] (2) A system that records personal data provided from multiple user terminals in a verifiable manner while protecting privacy, It is configured as a computer with a processor and memory, A data preprocessing unit performs differential privacy processing on the raw data received from the user terminal, adding noise based on a predetermined privacy budget (ε), to generate an anonymized feature vector that reduces personal identifiability while maintaining statistical characteristics. A data value calculation unit that uses a machine learning model to calculate the degree of influence that the anonymized feature vector has on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation unit that generates a zero-knowledge proof (ZKP) that proves the contribution score was calculated based on a valid algorithm without disclosing the raw data, The system includes a ledger linking unit that transmits the contribution score and the zero-knowledge proof as a transaction to the blockchain network, A data contribution verification system characterized in that a smart contract on the blockchain network executes a token issuance process corresponding to the contribution score only if the verification of the zero-knowledge proof is successful.

[0110] (2) Effects In this system, the data preprocessing unit performs differential privacy processing, adding noise based on a predetermined privacy budget (ε) to the raw data. This generates anonymized feature vectors with mathematically reduced personal identifiability, minimizing the risk of privacy infringement while maintaining the statistical properties necessary for AI model training and evaluation, thus enabling a balance between privacy and data utilization. Furthermore, the proof generation unit generates a zero-knowledge proof (ZKP) to mathematically prove that the contribution score was calculated based on the correct algorithm and data, without disclosing the raw data. This prevents the opacity (black box) of the contribution calculation process, which was a problem in conventional AI systems, and ensures the reliability of the entire system. In addition, the smart contract on the blockchain network executes the token issuance process only if the verification of this ZKP is successful. This configuration automatically eliminates fraudulent reward claims based on tampered scores and realizes fair value distribution without relying on a centralized administrator.

[0111] (3) The data contribution verification system of (2) is characterized in that the ledger linkage unit constructs a Merkle tree with the hash values ​​of multiple transactions that occurred within a certain period of time as leaf nodes, records only the root hash of the Merkle tree on the blockchain network, and has the function of generating and outputting a Merkle Proof indicating that a specific transaction is included in the root hash in response to an audit request from an external source.

[0112] (3) Effects (3) specifies a scalable configuration for the system in (2) to solve the challenges of blockchain network capacity limitations and processing costs (such as gas fees). The ledger linkage unit constructs a Merkle tree with the hash values ​​of multiple transactions (including contribution scores and zero-knowledge proofs) that occurred within a certain period as leaf nodes, and records only the root hash, which is the root of the tree, on the blockchain network. This configuration significantly reduces the cost of recording on-chain while maintaining data resistance to tampering, compared to recording all transaction details on-chain, and contributes to ensuring the scalability of the system. Furthermore, the system has a function to generate and output a Merkle proof that mathematically proves that a specific transaction is included in the root hash when an audit request is made from an external party. This allows external auditors and users to efficiently verify the legitimacy of their data processing records off-chain without having to verify the entire blockchain, thus achieving both transparency and system performance.

[0113] (4) The data value calculation unit is configured to take the anonymized feature vector as input and calculate the contribution score using the Shapley Value or gradient-based influence analysis. The data contribution verification system according to (2) or (3), characterized in that the data preprocessing unit has a function to dynamically adjust the privacy budget (ε) so that the amount of noise added in the differential privacy processing keeps the decrease in the accuracy of the score calculation by the data value calculation unit within a predetermined threshold.

[0114] (4) Effects (4) specifies the configuration for the systems of (2) or (3) to optimize the fairness of contribution calculation and the trade-off between privacy protection and data usefulness. First, the data valuation unit uses the Shapley Value or gradient-based impact analysis to enable the calculation of an accurate contribution score based not merely on the amount of data provided, but on the marginal contribution (data quality) to improving the accuracy of the machine learning model. This ensures that the rarity and quality of the data are appropriately reflected in the evaluation, resulting in a fair and satisfactory reward distribution for data providers. In addition, the data preprocessing unit has a function to dynamically adjust the privacy budget ($\epsilon$) so that the amount of noise added in differential privacy processing keeps the reduction in accuracy of the score calculation by the data valuation unit within a predetermined threshold. This dynamic adjustment prevents situations where privacy protection is strengthened too much, significantly impairing the usefulness of the data (effectiveness for AI learning), or conversely, where accuracy is prioritized too much, resulting in a violation of privacy, thus maintaining an optimal balance between privacy and usefulness in practical operation.

[0115] (5) (Token economics and inflation control) The data contribution verification system according to any one of (2) to (4), characterized in that, in the token issuance process, the smart contract or the ledger linkage unit has an inflation control logic that dynamically adjusts the issuance rate of new tokens based on the total amount of tokens issued, the current token demand index, and price feeds from external markets, and also has a burn mechanism that burns a portion of the tokens when it is determined that the market supply is excessive.

[0116] (5) Effects (5) specifies the configuration for supply control to maintain a healthy token economics in any of the systems described in (2) to (4). In this system, during the token issuance process, a smart contract or ledger linkage unit incorporates inflation control logic that dynamically adjusts the rate of new token issuance based on the total amount of issued tokens, the current token demand index, and price feeds from external markets. This mitigates the risk of excessive supply to the market and dilution of token value (inflation) caused by simply continuing to issue (mint) tokens, thereby stabilizing the economic value of the tokens. Furthermore, this system has a burn mechanism that burns a portion of the tokens if the market supply is deemed to be excessive, thereby reducing the supply and maintaining the scarcity of the tokens. These functions enable the automated programmatic control of the token supply to maintain the attractiveness (incentive) of the rewards in the long term, contributing to the sustainability of the ecosystem and the creation of a healthy economic zone.

[0117] (6) (Off-chain verification and oracles) The aforementioned ledger linkage unit communicates with verification nodes located outside the blockchain (off-chain) and transmits the results of verifying the integrity of large volumes of usage logs or highly confidential data to the smart contract via Oracle. The data contribution verification system according to any of (2) to (4), characterized in that the smart contract makes the approval signal from the oracle a mandatory condition for token distribution execution, in addition to the on-chain trigger conditions.

[0118] (6) Effects System (6) specifies a configuration for achieving both scalability and security in processing the large volume of usage logs generated in connection with the use of AI services in any of the systems (2) to (4). First, it adopts a configuration in which high-load computational processing, such as the analysis and integrity verification of large volumes of usage logs, is performed by verification nodes located off-chain. This avoids processing delays and high costs caused by performing all log processing on-chain, and improves the overall processing capacity (scalability) of the system. Next, the results of the integrity verification of logs or verification of highly confidential data obtained by the verification nodes off-chain are transmitted to the smart contract via an oracle. By using this oracle, even information located outside the blockchain can be incorporated as an execution condition (trigger) for the smart contract while ensuring reliability. As a result, the smart contract prevents distribution based on fraudulent logs and guarantees secure data exchange and fair value distribution by making the approval signal from the oracle a mandatory condition for token distribution execution, in addition to the on-chain trigger conditions.

[0119] (7) (Privacy settings UI and simulation) The data contribution verification system according to any one of (2) to (4) is characterized in that the data preprocessing unit provides a user interface that allows the user to select a differential privacy parameter (ε value) or anonymization level, dynamically changes the content of the preprocessing based on the setting value selected by the user, and includes a simulation function that predicts and displays in advance the impact that the selection will have on future token dividend amounts.

[0120] (7) Effects The system in (7) respects the user's right to self-determination regarding privacy settings, which is the privacy setting of the data provider, and defines a configuration to make the economic trade-offs associated with data provision transparent, in any of the systems in (2) to (4). The data preprocessing unit provides the user with a user interface that allows them to select differential privacy parameters (epsilon value) or anonymization levels. This allows users to adjust the settings according to their own privacy tolerance, encouraging their understanding and participation in the system. Furthermore, the system dynamically changes the content of preprocessing based on the settings selected by the user and includes a simulation function that predicts and displays in advance the impact that the selection of the privacy protection level will have on future token dividends through a decrease in data usefulness. This simulation function visualizes in advance the trade-off that enhanced privacy protection (high noise, low epsilon value) leads to a decrease in data usefulness and, consequently, a reduction in dividends, preventing users from suffering unintended disadvantages.

[0121] (8) (Multifaceted weighting calculation) The data value calculation unit includes a weighting model that calculates scores for data scarcity, update frequency, quality indicators, and usage context, rather than simply the quantity of data, and integrates these to calculate the contribution score, wherein the weighting model updates the weight coefficients by machine learning using past dividend performance and the effect of improving the accuracy of the AI ​​model as training data, making it one of the data contribution verification systems according to (2) to (4).

[0122] (8) Effects System (8) specifies a configuration for evaluating the qualitative contribution of data more fairly and from multiple perspectives in any of the systems (2) to (4). The data valuation unit has a weighting model that calculates scores for data scarcity, recency, quality indicators, and usage context, without depending on the simple amount of data provided, and integrates these to calculate a final contribution score. This multifaceted evaluation identifies data that truly contributed to improving the accuracy of the AI ​​model (data with a large marginal contribution), and by appropriately reflecting its scarcity and quality in the reward distribution, it realizes a fair and satisfactory valuation for data providers. Furthermore, the weighting model is configured to dynamically update the weight coefficients using machine learning with past dividend performance and the effect of improving the accuracy of the AI ​​model as training data. This makes it possible to automatically adapt to the optimal evaluation criteria even when the needs of the market or AI technology change, and to maintain the sustainability and fairness of the ecosystem in the long term.

[0123] (9) (Fraud detection and scheduler) The data contribution verification system according to any of (2) to (4) is characterized in that the smart contract or associated off-chain module comprises a scheduler that controls the frequency of dividend execution and a fraud detection module that performs statistical anomaly detection on transactions before dividend execution, and if an abnormal value or fraudulent operation pattern is detected, the dividend processing is automatically suspended and the system is branched to a re-verification flow.

[0124] (9) Effects System (9) specifies a configuration for any of the systems (2) through (4) to reduce the operational risk of token dividends and ensure flexibility in dividend timing. The smart contract or associated off-chain module includes a scheduler that controls the frequency of dividend execution, enabling the setting of dividend timing according to operational policies, such as daily or monthly. It also includes a fraud detection module that performs statistical anomaly detection on transactions before dividend execution. If this fraud detection module detects anomalies or fraudulent operation patterns, such as mass access by bots or fraudulent score manipulation through collusion (e.g., Sybil attacks), it automatically suspends dividend processing and branches to a re-verification flow. This prevents asset outflow due to erroneous dividends and significantly reduces the operational risk of the entire system.

[0125] (10) (Explainability · XAI) The data contribution verification system according to any one of (2) to (4) is characterized in that the data value calculation unit includes an explainability module that generates explanatory information indicating which features contributed to the calculated contribution score, and provides a dashboard on which users can view the explanatory information.

[0126] (10) Effects System (10) specifies a configuration for any of the systems (2) to (4) to enhance the transparency of contribution evaluation and ensure trust from data providers. The data valuation unit includes an explainability module to present the basis for the calculated contribution score to the user. This module generates explanatory information indicating which features (such as data rarity and quality) contributed positively to the contribution evaluation. Users can view this explanatory information through a dashboard that displays a breakdown and history of the contribution score. This configuration prevents the contribution calculation process from becoming an opaque black box, and by demonstrating that it is an evidence-based evaluation rather than an AI-based assessment, it alleviates user distrust of the system and contributes to improved reliability. Furthermore, because the breakdown of the evaluation is visualized, it also functions as evidence to facilitate objections in the event of calculation errors or dissatisfaction with the evaluation.

[0127] (11) (Governance vote) The system is a data contribution verification system according to any one of (2) to (4), characterized in that it has a governance function that allows token holders to vote on changes to the parameters of the inflation control logic or the evaluation criteria of the weighting model, and updates the parameters of the smart contract automatically or through a predetermined procedure based on the voting results.

[0128] (11) Effects The system in (11) incorporates a governance function in any of the systems in (2) to (4) that allows token holders to vote on changes to the parameters of the inflation control logic or the evaluation criteria of the weighting model. This distributes the authority to change important system parameters not only to operators but also to token holders, thereby realizing an autonomous ecosystem that is not controlled by any particular company. This makes it possible to flexibly change the rules based on community consensus, increasing adaptability to long-term environmental changes and ensuring the democratic operation of the system.

[0129] (12) (API Gateway and Normalization) A data contribution verification system according to any one of (2) to (4), comprising an API gateway having a normalization adapter that converts usage logs in different formats received from multiple external AI service providers into a standard format, wherein the API gateway verifies the digital signature and timestamp of the received logs before handing the data to the data value calculation unit.

[0130] (12) Effects The system in (12) ensures interoperability between heterogeneous systems and enables centralized processing by smart contracts by having an API gateway in any of the systems in (2) to (4) that has a normalization adapter that converts usage logs in different formats received from multiple external AI service providers into a standard format. Furthermore, the API gateway ensures the authenticity of the data by verifying the digital signature and timestamp of the received logs before handing over the data, thereby blocking log tampering and impersonation transmission at the entry point.

[0131] (13) (Receiving payment through My Number linkage) The aforementioned system is a data contribution verification system according to any of (2) to (4), characterized in that it includes a government ID linkage interface for connecting to a personal identification infrastructure operated by the government (including the My Number system), and the interface links the user's wallet address and government ID (e.g., My Number) in a privacy-protected manner using zero-knowledge proofs or hash functions, converts the tokens distributed by the smart contract into legal tender through a linked payment gateway, and automatically transfers them to a bank account such as a public funds receiving account linked to the government ID.

[0132] (13) Effects The system in (13) incorporates a government ID linkage interface in any of the systems in (2) to (4) to connect with a government-operated personal identification infrastructure. This links the user's wallet address and government ID in a privacy-protected manner using zero-knowledge proofs or hash functions, completing rigorous identity verification without providing unnecessary personal information to administrative agencies or system operators. Furthermore, smart contracts convert the distributed tokens into fiat currency through a linked payment gateway and automatically transfer them to bank accounts such as public fund receiving accounts. This prevents duplicate registrations by the same person, allows users to receive rewards in fiat currency without knowledge of cryptocurrency, and achieves both reliability and efficiency in distribution, as well as privacy protection.

[0133] (14) (Application to the Basic Income System) The data contribution verification system according to any of (2) to (4) is characterized in that the smart contract has a function to accumulate part or all of the revenue generated from the use of data collected from multiple users as a basic income fund pool, and has a basic income distribution logic that automatically distributes tokens as a basic living benefit from the pool on a regular and equal basis to all registered users who meet predetermined minimum data provision conditions, regardless of the data contribution of each individual user.

[0134] (14) Effects The system in (14) is one of the systems in (2) to (4) in which a smart contract has the function of accumulating part or all of the revenue generated from the use of data collected from multiple users as a basic income pool, and has a basic income distribution logic that automatically distributes tokens from the pool regularly and equally to all registered users who meet the predetermined minimum data provision conditions, regardless of the data contribution of individual users, thereby returning the revenue generated by AI to all ecosystem participants, not just some data providers, and functioning as a social safety net to prevent the widening of wealth inequality due to data disparity. In addition, the incentive of "receiving a minimum reward just for participating" encourages participation from a wide range of general users, and as a result has the effect of promoting the collection of diverse long-tail data which is essential for improving the generalization performance of AI models.

[0135] (15) (Detailed consent management and dynamic filtering) The aforementioned data preprocessing unit provides a consent management user interface on the user terminal that allows adjustment of the availability of each data category and the granularity of the data, and includes a consent history database that records the consent information set by the user, and a filtering logic that references the database in real time. The filtering logic is characterized by physically blocking or nullifying data in categories for which consent has not been obtained to prevent it from flowing into subsequent processing, and, if the user revokes their consent, sending a forget request to the data value calculation unit and related modules, which triggers the cessation of use of the user's past data and the exclusion of its reflection in the learning model, as a data contribution verification system according to any one of (2) to (4).

[0136] (15) Effects The system in (15), in any of the systems from (2) to (4), provides a consent management user interface on the user terminal that allows adjustment of not only the availability of data for each data category but also the granularity of the data. Unlike conventional exclusive consent settings, this allows users to flexibly control their privacy according to their own tolerance, reducing the psychological barrier to data provision. Furthermore, a filtering logic that references the consent history database in real time physically blocks or nullifies data in categories for which consent has not been obtained before it is passed to subsequent processing, thus reliably preventing unintended data leaks due to system configuration errors, etc. In addition, when a user withdraws their consent, the filtering logic sends a forgetting request to the data value calculation unit and related modules, triggering the cessation of use of the user's past data and exclusion from reflection in the learning model, thereby realizing the implementation of the "right to be forgotten" to enhance compliance with privacy protection regulations. Moreover, the mechanism by which the scope and granularity of this consent setting are linked to the data value calculation (reward amount) builds a healthy data ecosystem in which users themselves determine the balance between privacy risk and economic return.

[0137] (16) The data contribution verification system according to (12) is characterized in that the API gateway includes a security verification function that prevents replay attacks by performing digital signature verification using a public key and window checks of timestamps on usage logs having different data formats (JSON, XML, CSV, etc.) received from multiple different external AI service providers (Google, OpenAI, Microsoft, etc.), and a group of normalization adapters that absorb the differences in data formats for each provider and convert them into a standard log format within the system that includes a unique user identifier, data category, and transaction ID.

[0138] (16) Effects The system in (16) is an API gateway that, in the system in (12), has a set of normalization adapters that absorb the differences between providers and convert usage logs, which have different data formats and are received from multiple external AI service providers of different origins, into a standard log format within the system. This ensures a neutral platform that is not dependent on any particular vendor and interoperability between heterogeneous systems, enabling centralized processing by smart contracts. Furthermore, by having a security verification function that performs digital signature verification using public keys and window checks of timestamps, it verifies the authenticity of the sender and prevents fraudulent reward claims (replay attacks) by resending past logs at the source, contributing to the prevention of fraudulent dividends. This architecture can accommodate the entry of new AI service providers in the future by simply adding the corresponding new adapters, dramatically improving the scalability and availability of the system.

[0139] (17) (Macroeconomically linked liquidity supply) The data contribution verification system (14) is characterized in that the basic income fund pool is equipped with a macroeconomic linkage interface that connects to an external national AI fund, a robot tax collection system, or a central bank digital currency (CBDC) system, and the interface has the function of detecting the growth in gross domestic product (GDP) or excess profits of AI companies resulting from automation by AI or robots, and automatically tokenizing these and continuously injecting them into the fund pool as liquidity.

[0140] (17) Effects System (17), in conjunction with System (14), innovatively expands the concept of basic income (BI). While conventional BI aims to guarantee a minimum standard of living and has limited funding sources, this system significantly strengthens the fundraising mechanism for the basic income pool.

[0141] Specifically, the basic income fund pool will not only depend on data usage fees within the system, i.e., compensation based on the value of data used by information systems such as AI, but will also be able to comprehensively incorporate a portion of the added value created by national-scale economic growth.

[0142] This expanded funding mechanism will go beyond the framework of basic income as a mere minimum living guarantee, making it possible to secure an overwhelming amount of funding to realize Universal High Income (UHI) for every citizen—a generous and stable benefit far exceeding that of conventional BI. As a result, it is expected that citizens will be freed from economic anxiety and be able to dedicate themselves to more creative and socially beneficial activities.

[0143] This UHI establishes the foundation for a future economic system in which data is recognized as a new factor of production and its value is fairly distributed to society as a whole. It forms the core of an innovative socio-economic model that aims to broadly share the wealth generated by AI and data, correct economic inequality, and improve the overall well-being of society.

[0144] (18) (Purchasing Power Parity Adjustment and Real Living Security) The data contribution verification system according to (14) or (17), characterized in that the smart contract works in conjunction with a purchasing power parity (PPP) adjustment oracle that monitors in real time the market prices of a basket of essential goods including energy prices, food prices, and housing costs, and the basic income dividend logic dynamically calculates and adjusts the amount of tokens paid per person necessary to maintain real purchasing power in accordance with deflation or inflation, based on a real cost of living index obtained from the oracle.

[0145] (18) Effects The system in (18) is positioned as an important function that extends the AI-based data contribution verification system presented in (14) or (17). The primary purpose of this function is to address the potential risks that advances in AI technology pose to socioeconomics, particularly the possibility that increased productivity could trigger extreme price declines (hyperdeflation).

[0146] As AI takes over many aspects of labor and production, the production costs of goods and services will dramatically decrease, and their quality will improve, making hyperdeflation, a sustained and significant decline in prices across the market, more likely. As a result, even if nominal income and benefit levels are maintained, there is a risk that corporate profitability and the value of individual savings will be impaired.

[0147] While conventional reward and benefit systems focused on "nominal benefit amounts," system (18) focuses on accurately evaluating and maintaining the real purchasing power of rewards and benefits provided to users, even in price fluctuations such as hyperdeflation. Specifically, it has a mechanism to dynamically adjust nominal benefit amounts by linking them to price indices (such as the Consumer Price Index and the AI ​​Deflation Adjustment Index).

[0148] This function of maintaining effective purchasing power ensures that the benefits of the AI-driven productivity revolution are not merely technological advancements, but rather guarantee that users' living standards remain constant or improve over time. Even with price fluctuations, people's ability to purchase necessary goods and services—that is, their quality of life—is protected, which is expected to curb social unrest and the widening of economic inequality. Furthermore, by ensuring that the rewards for data contributions to AI do not easily lose value due to overall economic fluctuations, a sense of fairness among system participants (data providers and contributors) is guaranteed, creating a foundation for increased reliability in long-term system use.

[0149] (19) (Autonomous governance with censorship resistance) The data contribution verification system of any of (14) to (19) is characterized in that the smart contract and the basic income distribution logic are deployed as immutable code on the blockchain, have censorship resistance functions that reject arbitrary suspension of payments or freezing of accounts at the program level by specific administrators, government agencies, or corporations, and changes to the system parameters are carried out only by consensus through decentralized voting (DAO) based on the amount of the tokens and reputation tokens held.

[0150] (19) Effects The system in (19) is one of the systems in (14) through (18), in which the smart contract and basic income distribution logic are deployed as immutable code on the blockchain and have censorship-resistant features that reject arbitrary suspension of payments or freezing of accounts at the program level by specific administrators, government agencies, or corporations. This protects the basic benefit system from arbitrary manipulation and political pressure by specific central authorities, ensuring its reliability as a permanent safety net. Furthermore, the configuration in which changes to the system parameters are carried out only by consensus through decentralized voting (DAO) based on the amount of the aforementioned tokens and reputation tokens held ensures a high degree of autonomy and democratic operation in the decision-making process regarding wealth distribution, while preventing centralized control.

[0151] (20) A system for distributing value derived from production activities by artificial intelligence or robots to multiple users using smart contracts deployed on a blockchain network, comprising: a macroeconomic interface that communicates with an external national AI fund, tax system, or central bank digital currency system to detect and tokenize the growth in gross domestic product (GDP) or profits of AI-related companies associated with the spread of AI, and supplies them as liquidity to a basic income pool; a purchasing power parity (PPP) adjusted oracle that monitors the market prices of a basket of necessities, including energy, food, and housing costs, in real time and obtains a real cost of living index that shows the fluctuations in deflation or inflation associated with the reduction in production costs due to AI; and a dividend logic that uses tokens accumulated in the basic income pool as the source, dynamically calculates the amount of per capita benefit necessary to maintain real purchasing power based on the real cost of living index obtained from the purchasing power parity adjusted oracle, and automatically distributes it to the user's wallet address.

[0152] (20) Effects The system in (20) communicates with external national AI funds, tax systems, or central bank digital currency systems and includes a macroeconomic linkage interface (gateway function) that detects and tokenizes the growth in gross domestic product (GDP) or profits of AI-related companies associated with the spread of AI, and supplies it as liquidity to the basic income fund pool, thereby securing the financial resources necessary to realize a universal high income that surpasses conventional basic income. Furthermore, it includes a purchasing power parity (PPP) adjusted oracle that monitors the market prices of a basket of necessities, including energy, food, and housing costs, in real time and obtains a real cost of living index that shows fluctuations in deflation or inflation, and a dividend logic that uses tokens accumulated in the fund pool as the source and dynamically calculates the amount of per capita benefit necessary to maintain real purchasing power based on the real cost of living index obtained from the oracle, and automatically distributes it to the user's wallet address, thereby enabling it to respond to rapid price fluctuations associated with productivity improvements by AI and maintain and improve real living standards rather than nominal amounts.

[0153] (twenty one) The AI-derived value distribution system as described in (20), further comprising a community contribution certification protocol that records mutual evaluations of care work, creative activities, or volunteer activities among users, independently of monetary value exchange, wherein the protocol issues non-transferable reputation tokens (Soulbound Tokens) based on actions of gratitude and approval from others, and accumulates these in an individual's social capital ledger, thereby visualizing an individual's social credit in a society where work is no longer an obligation.

[0154] (21) Effects The system in (21) incorporates a community contribution certification protocol that records mutual evaluations of care work, creative activities, or volunteer activities among users, independently of monetary value exchange, thereby making visible activities that are difficult to evaluate by market principles in a society where work is no longer an obligation, in the form of gratitude and recognition from others.

[0155] Specifically, this protocol first involves users mutually sending and securely recording digitally signed transactions containing evaluation tags and weighted values ​​indicating "gratitude" or "approval" for care work (e.g., helping people in need), creative activities, or volunteer work performed by others. Next, based on the recorded evaluations, non-transferable reputation tokens (SBT: Soulbound Token) are issued and accumulated in the individual's social capital ledger.

[0156] Because SBT cannot be bought, sold, or transferred, it purely embodies an individual's social activity history and creditworthiness. This provides a new measure of value other than money—"social credit"—and incentivizes people to continue contributing to society voluntarily. Furthermore, it makes it easier for people outside the framework of market competition to find a "place" and "recognition" within their community, preventing social isolation and contributing to the construction of resilient social capital that encompasses diverse talents and activities.

[0157] (twenty two) The aforementioned dividend logic is deployed as immutable code on the blockchain and has censorship-resistant functionality that rejects arbitrary suspension of payments or freezing of accounts by specific administrators, government agencies, or corporations at the program level, as an AI-derived value distribution system according to (20) or (21).

[0158] (22) Effects The system in (22) protects the basic livelihood benefit system from arbitrary manipulation by specific central authorities and political pressure, by having a censorship-resistant function that rejects arbitrary suspension of benefits or freezing of accounts at the program level by specific administrators, government agencies, or corporations, in the system in (20) or (21), and by ensuring reliability as a permanent safety net through automated execution by program and decentralized management.

[0159] (twenty three) A method for recording the contribution of data in a verifiable manner, in order to achieve a fair distribution of value created by AI based on data provided from multiple user terminals, A computer equipped with a processor and memory, A data preprocessing step involves performing privacy protection processing on the data received from the user terminal to generate anonymized features, A data value calculation step that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation step that generates proof information to prove that the calculation of the contribution score was performed based on a valid algorithm, A data contribution verification method characterized by performing a ledger linkage step of transmitting the contribution score and the proof information to a distributed ledger on which the token issuance process is performed.

[0160] (23) Effects This data contribution verification method ensures data anonymity while maintaining the characteristics necessary for AI model training and reducing the risk of privacy infringement by performing a data preprocessing step on data received from user terminals to generate anonymized features with reduced personal identifiability through privacy protection processing. Next, a data value calculation step quantitatively calculates the degree to which the anonymized features have an impact on improving the accuracy of the training model using a machine learning model, and outputs the result as a contribution score. This enables objective evaluation based on qualitative contribution rather than quantitative contribution, providing data providers with a basis for granting fair compensation commensurate with their contribution. Furthermore, a proof generation step generates proof information that proves the calculation of the contribution score was performed based on a legitimate algorithm, and a ledger linkage step transmits the contribution score and the proof information to a distributed ledger where token issuance processing is performed. This prevents the score calculation process from becoming a black box, allows for verification of the validity of the calculation without relying on a central administrator or third-party organization, and ensures the reliability and transparency of the entire system.

[0161] (twenty four) A method for recording data contributions from multiple user terminals in a verifiable manner while protecting privacy, based on personal data provided from multiple user terminals, A computer equipped with a processor and memory, A data preprocessing step involves performing differential privacy processing on the raw data received from the user terminal, adding noise based on a predetermined privacy budget (ε), to generate an anonymized feature vector with reduced personal identifiability while maintaining statistical characteristics. A data value calculation step that uses a machine learning model to calculate the degree of influence that the anonymized feature vector has on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation step that generates a zero-knowledge proof (ZKP) that proves the contribution score was calculated based on a valid algorithm without disclosing the raw data, A data contribution verification method characterized by performing a ledger linkage step which involves sending the contribution score and the zero-knowledge proof as a transaction to the blockchain network, and executing a token issuance process on the condition that the verification of the zero-knowledge proof by a smart contract on the blockchain network is successful.

[0162] (24) Effects This data contribution verification method generates anonymized feature vectors with mathematically reduced personal identifiability by performing differential privacy processing in the data preprocessing step, which adds noise based on a predetermined privacy budget (ε) to the raw data. This minimizes the risk of privacy infringement while maintaining the statistical properties necessary for training and evaluating AI models, thus enabling a balance between privacy and data utilization. Furthermore, in the proof generation step, a zero-knowledge proof (ZKP) is generated to mathematically prove that the contribution score was calculated based on the correct algorithm and data, without disclosing the raw data. This prevents the lack of transparency (black boxing) in the contribution calculation process, which was a problem in conventional AI systems, and ensures the reliability of the entire system. In addition, the ledger linkage step ensures that the smart contract on the blockchain network only executes the token issuance process if the ZKP is successfully verified. This automatically eliminates fraudulent reward claims based on tampered scores and realizes fair value distribution without relying on a centralized administrator.

[0163] (twenty five) A program to cause a computer to function as a data contribution verification system that records the contribution of personal data provided from multiple user terminals in a verifiable manner in order to achieve a fair distribution of value created by AI, The aforementioned computer, A data preprocessor performs privacy protection processing on the data received from the user terminal to generate anonymized features. A data value calculation unit that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score. A proof generation unit that generates proof information to prove that the aforementioned contribution score was calculated based on a valid algorithm, and A data contribution verification program that functions as a ledger linkage unit, transmitting the aforementioned contribution score and the aforementioned proof information to a distributed ledger where the token issuance process is performed.

[0164] (25) Effects This data contribution verification program utilizes a computer as a data preprocessing unit to perform privacy protection processing on data received from user terminals, generating anonymized features with reduced personal identifiability. This ensures data anonymity while maintaining the characteristics necessary for AI model training, thereby providing a data provision platform that reduces the risk of privacy infringement. Furthermore, by utilizing the computer as a data value calculation unit, it quantitatively calculates the impact of the anonymized features on improving the accuracy of machine learning models and outputs the result as a contribution score. This objectively evaluates value based on qualitative contribution rather than quantitative contribution, providing a basis for granting fair compensation to data providers commensurate with their contributions. Moreover, by utilizing the computer as a proof generation unit and ledger linkage unit, it generates proof information that mathematically and cryptographically guarantees the validity of the calculated contribution score based on a correct algorithm. This is recorded in a distributed ledger, preventing the score calculation process from becoming an opaque black box. This allows for verification of the calculation's validity without relying on a central administrator or third-party organization, ensuring the reliability and transparency of the entire system.

[0165] (26) A program for causing a computer to function as a data contribution verification system that records the contribution of data in a verifiable manner while protecting privacy, based on personal data provided from multiple user terminals, The aforementioned computer, A data preprocessor performs differential privacy processing on the raw data received from the user terminal, adding noise based on a predetermined privacy budget (ε), to generate an anonymized feature vector that reduces personal identifiability while maintaining statistical characteristics. A data value calculation unit that uses a machine learning model to calculate the degree of influence that the anonymized feature vector has on improving the accuracy of the learning model, and outputs the calculation result as a contribution score. A proof generation unit that generates a zero-knowledge proof (ZKP) that proves the contribution score was calculated based on a valid algorithm without disclosing the raw data, and A data contribution verification program that functions as a ledger linkage unit, which transmits the contribution score and the zero-knowledge proof as a transaction to the blockchain network and executes a token issuance process conditional on the successful verification of the zero-knowledge proof by a smart contract on the blockchain network.

[0166] (26) Effects This data contribution verification program utilizes a computer as a data preprocessing unit to perform differential privacy processing, adding noise based on a predetermined privacy budget (ε) to raw data. This generates anonymized feature vectors with mathematically reduced personal identifiability, minimizing the risk of privacy infringement while maintaining the statistical properties necessary for AI model training, thus enabling a balance between privacy and data utilization. Furthermore, by utilizing the computer as a proof generation unit, it generates zero-knowledge proofs (ZKPs) to mathematically prove that the contribution score was calculated based on the correct algorithm and data, without disclosing the raw data. This prevents the opacity (black boxing) of the contribution calculation process that has been a problem in conventional AI systems, ensuring the reliability of the entire system. In addition, by utilizing the computer as a ledger linkage unit, it transmits the contribution score and the zero-knowledge proof to the blockchain network. Token issuance is triggered only upon successful verification of the ZKP by a smart contract on the network. This automatically eliminates fraudulent reward claims based on tampered scores, achieving fair value distribution without relying on a centralized administrator.

[0167] (27) A method of distributing value derived from production activities by artificial intelligence or robots to multiple users using smart contracts deployed on a blockchain network, A macroeconomic linkage step involves a computer equipped with a processor and memory communicating with an external national AI fund, tax system, or central bank digital currency system to detect and tokenize the GDP growth or profits of AI-related companies resulting from the proliferation of AI, and supplying them as liquidity to a basic income pool. A purchasing power parity (PPP) adjustment step that monitors market prices in real time for a basket of essential goods, including energy, food, and housing costs, and obtains a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions, An AI-derived value distribution method characterized by including a dividend execution step of dynamically calculating the amount of benefits per person necessary to maintain real purchasing power based on the acquired real cost of living index, using tokens accumulated in the basic income fund pool as the source, and automatically distributing it to the user's wallet address.

[0168] (27) Effects This AI-derived value distribution method communicates with external national AI funds, tax systems, or central bank digital currency systems, and includes a macroeconomic linkage step that detects the growth in Gross Domestic Product (GDP) or profits of AI-related companies derived from AI and robot production activities and supplies them as tokenized liquidity to the basic income pool. This ensures that the financial resources for realizing Universal High Income (UHI) funded by national wealth are not solely dependent on data usage fees. Furthermore, it includes a purchasing power parity adjustment step that monitors the market prices of a basket of necessities, including energy, food, and housing costs, in real time and obtains a real cost of living index that shows price fluctuations. Based on the obtained real cost of living index, it dynamically calculates the per capita benefit amount necessary to maintain real purchasing power and automatically distributes it to the user's wallet address. This allows the system to adapt to deflationary or inflationary fluctuations associated with the spread of AI, enabling it to maintain or improve the user's real standard of living rather than just a nominal amount.

[0169] (28) A program that enables a computer to function as an AI-derived value distribution system that uses smart contracts deployed on a blockchain network to distribute value derived from production activities by artificial intelligence or robots to multiple users, To the aforementioned computer, A macroeconomic linkage function that communicates with external national AI funds, tax systems, or central bank digital currency systems to detect and tokenize the GDP growth or profits of AI-related companies resulting from the spread of AI, and supplies them as liquidity to the basic income pool. A purchasing power parity (PPP) adjustment function that monitors market prices in real time for a basket of essential goods, including energy, food, and housing costs, and obtains a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions, and An AI-derived value distribution program to implement a dividend logic function that uses tokens accumulated in the aforementioned basic income fund pool as the source of funds, dynamically calculates the amount of benefit per person necessary to maintain real purchasing power based on the acquired real cost of living index, and automatically distributes it to the user's wallet address.

[0170] (28) Effects This AI-derived value distribution program enables a macroeconomic linkage function that allows computers to communicate with external national AI funds, tax systems, or central bank digital currency systems to detect and tokenize the growth in gross domestic product (GDP) or profits of AI-related companies associated with the spread of AI, and supply them as liquidity to the basic income fund pool. This ensures that the financial resources for realizing Universal High Income (UHI), which is funded by national-scale economic growth, are not solely dependent on internal system profits. Furthermore, it implements a purchasing power parity adjustment function that monitors the market prices of a basket of necessities, including energy, food, and housing costs, in real time, and obtains a real cost of living index that shows fluctuations in deflation or inflation due to the reduction in production costs caused by AI. It also implements a dividend logic function that uses tokens accumulated in the fund pool as the source of funds, dynamically calculates the amount of per capita benefit necessary to maintain real purchasing power based on the obtained real cost of living index, and automatically distributes it to the user's wallet address. This allows the program to adapt to rapid price fluctuations associated with the spread of AI and maintain or improve the real standard of living of users, rather than just nominal amounts. [Explanation of Symbols]

[0171] 10 Data Preprocessing Section 15. Filtering Logic 16 Consent History Database 20 Data Valuation Department 30 Proof Generation Unit 40 Ledger Linkage Department 60 API Gateways 70 Off-chain Verification Nodes 100 Data Contribution Verification System 200 user terminals 220 Consent Management UI 300 External AI Service Providers 400 Smart Contracts 410 Inflation Control Logic 420 Burn mechanism 430 Dividend Scheduler 440 Fraud Detection Modules 460 Basic Income Fund Pool 470 Basic Income Dividend Logic 500 Blockchain Networks 510 Oracle 520 Merkle Tree 600 Government ID Linkage Interface 800 Payment Gateways

Claims

1. A system that records the contribution of data in a verifiable manner in order to achieve a fair distribution of value created by AI based on data provided from multiple user terminals, It is configured as a computer with a processor and memory, A data preprocessing unit performs privacy protection processing on data received from the user terminal to generate anonymized features, A data value calculation unit that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation unit that generates proof information to prove that the aforementioned contribution score was calculated based on a valid algorithm, A data contribution verification system characterized by comprising: a ledger linkage unit that transmits the contribution score and the proof information to a distributed ledger on which token issuance processing is performed.

2. A system that records personal data provided from multiple user terminals in a verifiable manner while protecting privacy, It is configured as a computer with a processor and memory, A data preprocessing unit performs differential privacy processing on the raw data received from the user terminal, adding noise based on a predetermined privacy budget, to generate an anonymized feature vector that reduces personal identification while maintaining statistical characteristics. A data value calculation unit that uses a machine learning model to calculate the degree of influence that the anonymized feature vector has on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation unit that generates a zero-knowledge proof that proves the contribution score was calculated based on a valid algorithm without disclosing the raw data, The system includes a ledger linking unit that transmits the contribution score and the zero-knowledge proof as a transaction to the blockchain network, A data contribution verification system characterized in that a smart contract on the blockchain network executes a token issuance process corresponding to the contribution score only if the verification of the zero-knowledge proof is successful.

3. The data contribution verification system according to claim 2, characterized in that the ledger linkage unit constructs a Merkle tree with the hash values ​​of multiple transactions that occurred within a certain period of time as leaf nodes, records only the root hash of the Merkle tree in the blockchain network, and has a function to generate and output a Merkle certificate indicating that a specific transaction is included in the root hash in response to an audit request from an external source.

4. The data value calculation unit is configured to take the anonymized feature vector as input and calculate the contribution score using Shapley values ​​or gradient-based influence analysis. The data contribution verification system according to claim 2, characterized in that the data preprocessing unit has a function to dynamically adjust the privacy budget so that the amount of noise added in the differential privacy processing keeps the reduction in the accuracy of the score calculation by the data value calculation unit within a predetermined threshold.

5. The data contribution verification system according to any one of claims 2 to 4, characterized in that, in the token issuance process, the smart contract or the ledger linkage unit includes an inflation control logic that dynamically adjusts the issuance rate of new tokens based on the total amount of tokens issued, the current token demand index, and price feeds from external markets, and also includes a burn mechanism that burns a portion of the tokens when it is determined that the market supply is excessive.

6. The aforementioned ledger linkage unit communicates with a verification node located outside the blockchain and transmits the results of verifying the integrity of large volumes of usage logs or highly confidential data to the smart contract via an oracle. The data contribution verification system according to any one of claims 2 to 4, characterized in that the smart contract makes the approval signal from the oracle a mandatory condition for token distribution execution, in addition to the on-chain trigger conditions.

7. The data preprocessing unit provides a user interface that allows the user to select differential privacy parameters or anonymization levels, dynamically changes the content of the preprocessing based on the settings selected by the user, and includes a simulation function that predicts and displays in advance the impact of the selection on future token dividend amounts, thus providing a data contribution verification system according to any one of claims 2 to 4.

8. The data value calculation unit comprises a weighting model that calculates scores for data rarity, update frequency, quality indicators, and usage context, rather than simply the quantity of data, and integrates these to calculate the contribution score, wherein the weighting model updates the weight coefficients by machine learning using past dividend performance and the accuracy improvement effect of the AI ​​model as training data, as described in any one of claims 2 to 4.

9. The data contribution verification system according to any one of claims 2 to 4, wherein the smart contract or associated off-chain module comprises a scheduler that controls the frequency of dividend execution and a fraud detection module that performs statistical anomaly detection on transactions before dividend execution, and if an abnormal value or fraudulent operation pattern is detected, the system automatically suspends dividend processing and branches to a re-verification flow.

10. The data contribution verification system according to any one of claims 2 to 4, wherein the data value calculation unit includes an explainability module that generates explanatory information indicating which features contributed to the calculated contribution score, and provides a dashboard on which users can view the explanatory information.

11. The data contribution verification system according to any one of claims 2 to 4, characterized in that the system includes a governance function that allows token holders to vote on changes to the parameters of the inflation control logic or the evaluation criteria of the weighting model, and updates the parameters of the smart contract automatically or through a predetermined procedure based on the voting results.

12. A data contribution verification system according to any one of claims 2 to 4, comprising an API gateway having a normalization adapter that converts usage logs in different formats received from multiple external AI service providers into a standard format, wherein the API gateway verifies the electronic signature and timestamp of the received logs before handing the data to the data value calculation unit.

13. The data contribution verification system according to any one of claims 2 to 4, characterized in that the system includes a government ID linkage interface for connecting to a personal identification infrastructure operated by the government, the interface links the user's wallet address and the government ID in a privacy-protected manner using zero-knowledge proofs or hash functions, converts the tokens distributed by the smart contract into legal tender through a linked payment gateway, and automatically transfers them to a bank account such as a public funds receiving account linked to the government ID.

14. The data contribution verification system according to any one of claims 2 to 4, characterized in that the smart contract has a function to accumulate part or all of the revenue generated from the use of data collected from multiple users as a basic income fund pool, and has a basic income distribution logic that automatically distributes tokens as a basic living benefit from the pool on a regular and equal basis to all registered users who meet predetermined minimum data provision conditions, regardless of the data contribution of each individual user.

15. The data preprocessing unit provides a consent management user interface on the user terminal that allows adjustment of the availability of each data category and the granularity of the data, and includes a consent history database that records the consent information set by the user, and a filtering logic that references the database in real time. The filtering logic physically blocks or nullifies data in categories for which consent has not been obtained to prevent it from flowing into subsequent processing, and if the user withdraws their consent, it sends a forgetting request to the data value calculation unit and related modules, triggering the cessation of use of the user's past data and the exclusion of its reflection in the learning model, as described in any one of claims 2 to 4.

16. The data contribution verification system according to claim 12, characterized in that the API gateway includes a security verification function that prevents replay attacks by performing digital signature verification using a public key and a timestamp window check on usage logs having different data formats received from multiple external AI service providers that are different from each other, and a group of normalization adapters that absorb the differences in data formats for each provider and convert them into a standard log format within the system that includes a unique user identifier, data category, and transaction ID.

17. The data contribution verification system according to claim 14, wherein the basic income fund pool is equipped with a macroeconomic linkage interface that connects to an external national AI fund, a robot tax collection system, or a central bank digital currency system, and the interface has the function of detecting the growth in gross domestic product or excess profits of AI companies resulting from automation by AI or robots, and automatically tokenizing these and continuously injecting them into the fund pool as liquidity.

18. The data contribution verification system according to claim 14, characterized in that the smart contract works in conjunction with a purchasing power parity adjustment oracle that monitors in real time the market prices of a basket of essential goods, including energy prices, food prices, and housing costs, and the basic income dividend logic dynamically calculates and adjusts the amount of tokens paid per person necessary to maintain real purchasing power in accordance with deflation or inflation, based on a real cost of living index obtained from the oracle.

19. The data contribution verification system according to claim 14, characterized in that the smart contract and the basic income distribution logic are deployed as immutable code on the blockchain, have censorship-resistant features that programmatically reject arbitrary suspension of payments or freezing of accounts by specific administrators, government agencies, or corporations, and that changes to the system parameters are carried out only by consensus through decentralized voting based on the amount of tokens and reputation tokens held.

20. A system that uses smart contracts deployed on a blockchain network to distribute value derived from production activities by artificial intelligence or robots to multiple users, A macroeconomic interface that communicates with external national AI funds, tax systems, or central bank digital currency systems to detect and tokenize the growth in GDP or profits of AI-related companies resulting from the spread of AI, and supplies them as liquidity to the basic income pool. A purchasing power parity-adjusted oracle that monitors market prices in real time for a basket of essential goods, including energy, food, and housing costs, and obtains a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions, An AI-derived value distribution system characterized by comprising: a dividend logic that uses tokens accumulated in the basic income fund pool as its source, dynamically calculates the amount of benefits per person necessary to maintain real purchasing power based on the real cost of living index obtained from the purchasing power parity adjusted oracle, and automatically distributes it to the user's wallet address.

21. The AI-derived value distribution system according to claim 20, further comprising a community contribution certification protocol that records mutual evaluations of care work, creative activities, or volunteer activities among users, independently of monetary value exchange, wherein the protocol issues non-transferable reputation tokens based on actions of gratitude and approval from others, and accumulates these in an individual's social capital ledger, thereby making an individual's social credit visible in a society where work is no longer an obligation.

22. The AI-derived value distribution system according to 20 or 21, characterized in that the dividend logic is deployed as immutable code on the blockchain and has censorship-resistant functionality that programmatically rejects arbitrary suspension of payments or freezing of accounts by specific administrators, government agencies, or corporations.

23. A method for recording the contribution of data in a verifiable manner, in order to achieve a fair distribution of value created by AI based on data provided from multiple user terminals, A computer equipped with a processor and memory, A data preprocessing step involves performing privacy protection processing on the data received from the user terminal to generate anonymized features, A data value calculation step that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation step that generates proof information to prove that the calculation of the contribution score was performed based on a valid algorithm, A data contribution verification method characterized by performing a ledger linkage step of transmitting the contribution score and the proof information to a distributed ledger on which the token issuance process is performed.

24. A method for recording personal data provided from multiple user terminals in a way that allows for verification of the data's contribution while protecting privacy, A computer equipped with a processor and memory, A data preprocessing step involves performing differential privacy processing on the raw data received from the user terminal, adding noise based on a predetermined privacy budget, to generate an anonymized feature vector that reduces personal identifiability while maintaining statistical characteristics. A data value calculation step that uses a machine learning model to calculate the degree of influence that the anonymized feature vector has on improving the accuracy of the learning model, and outputs the calculation result as a contribution score, A proof generation step that generates a zero-knowledge proof that proves the contribution score was calculated based on a valid algorithm without disclosing the raw data, A data contribution verification method characterized by performing a ledger linkage step which involves sending the contribution score and the zero-knowledge proof as a transaction to the blockchain network, and executing a token issuance process on the condition that the verification of the zero-knowledge proof by a smart contract on the blockchain network is successful.

25. A program to cause a computer to function as a data contribution verification system that records the contribution of personal data provided from multiple user terminals in a verifiable manner in order to achieve a fair distribution of value created by AI, The aforementioned computer, A data preprocessor performs privacy protection processing on the data received from the user terminal to generate anonymized features. A data value calculation unit that uses a machine learning model to calculate the degree to which the anonymized features have an impact on improving the accuracy of the learning model, and outputs the calculation result as a contribution score. A proof generation unit that generates proof information to prove that the aforementioned contribution score was calculated based on a valid algorithm, and A data contribution verification program that functions as a ledger linkage unit, transmitting the aforementioned contribution score and the aforementioned proof information to a distributed ledger where the token issuance process is performed.

26. A program for causing a computer to function as a data contribution verification system that records the contribution of data in a verifiable manner while protecting privacy, based on personal data provided from multiple user terminals, The aforementioned computer, A data preprocessor performs differential privacy processing on the raw data received from the user terminal, adding noise based on a predetermined privacy budget, to generate an anonymized feature vector that reduces personal identifiability while maintaining statistical characteristics. A data value calculation unit that uses a machine learning model to calculate the degree of influence that the anonymized feature vector has on improving the accuracy of the learning model, and outputs the calculation result as a contribution score. A proof generation unit that generates a zero-knowledge proof that proves the contribution score was calculated based on a valid algorithm without disclosing the raw data, and A data contribution verification program that functions as a ledger linkage unit, which transmits the contribution score and the zero-knowledge proof as a transaction to the blockchain network and executes a token issuance process conditional on the successful verification of the zero-knowledge proof by a smart contract on the blockchain network.

27. A method of distributing value derived from production activities by artificial intelligence or robots to multiple users using smart contracts deployed on a blockchain network, A macroeconomic linkage step involves a computer equipped with a processor and memory communicating with an external national AI fund, tax system, or central bank digital currency system to detect and tokenize the growth in GDP or profits of AI-related companies resulting from the proliferation of AI, and supplying them as liquidity to a basic income pool. A purchasing power parity adjustment step involves monitoring market prices in real time for a basket of essential goods, including energy, food, and housing costs, and obtaining a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions. An AI-derived value distribution method characterized by including a dividend execution step of dynamically calculating the amount of per capita benefit necessary to maintain real purchasing power based on the acquired real cost of living index, using tokens accumulated in the basic income fund pool as the source, and automatically distributing it to the user's wallet address.

28. A program that enables a computer to function as an AI-derived value distribution system that uses smart contracts deployed on a blockchain network to distribute value derived from production activities by artificial intelligence or robots to multiple users, To the aforementioned computer, A macroeconomic linkage function that communicates with external national AI funds, tax systems, or central bank digital currency systems to detect and tokenize the growth in GDP or profits of AI-related companies resulting from the spread of AI, and supplies them as liquidity to the basic income pool. A purchasing power parity adjustment function that monitors market prices in real time for a basket of essential goods, including energy, food, and housing costs, and obtains a real cost of living index that shows deflationary or inflationary fluctuations due to AI-driven production cost reductions, and An AI-derived value distribution program to implement a dividend logic function that uses tokens accumulated in the aforementioned basic income fund pool as the source of funds, dynamically calculates the amount of benefits per person necessary to maintain real purchasing power based on the acquired real cost of living index, and automatically distributes it to the user's wallet address.

Images