AI Dataset Provenance Tracking With Blockchain NFT Access Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to securely protect the provenance, ownership, and integrity of artificial intelligence (AI) datasets and models, leading to potential biases, security vulnerabilities, and compliance issues in enterprise environments.
Innovation Solution
Represent AI datasets and models as non-fungible tokens (NFTs) on a blockchain, associating them with private keys and biometric authentication for enhanced security and transparency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If AI datasets and models are stored using traditional methods, then accessibility and ease of use are improved, but security, provenance tracking, and ownership protection deteriorate
Solution Approach 1:
The patent introduces blockchain technology as an intermediary layer between AI datasets/models and users. The blockchain serves as a trusted mediator that records provenance, ownership, and access rights immutably, while traditional storage systems maintain accessibility. Smart contracts act as automated intermediaries that enforce access policies and track usage, resolving the contradiction by providing security through the intermediary blockchain infrastructure without compromising user accessibility to the actual AI assets.
Solution Approach 2:
The patent segments the AI asset management system into distinct components: the actual AI datasets/models stored in traditional accessible locations, blockchain-based provenance and ownership records, and smart contract-based access control mechanisms. This segmentation allows each component to fulfill its specific function - traditional storage provides accessibility, blockchain provides immutable provenance tracking, and smart contracts provide automated security enforcement - thereby resolving the contradiction between ease of operation and reliability.
2Productivity
If AI datasets are made publicly accessible for training purposes, then utility and productivity are improved, but data privacy and security deteriorate
Solution Approach 1:
The patent applies local quality by implementing different access permissions and security levels for different portions of AI datasets. Sensitive portions of data can be encrypted or restricted to specific authorized users, while non-sensitive portions remain publicly accessible for training. This allows the system to provide high utility for general training purposes while maintaining data privacy and security for sensitive information through localized differential access controls.
Solution Approach 2:
The blockchain and smart contracts serve as intermediaries that enable controlled public access to AI datasets. The smart contracts verify user credentials, enforce access policies, and track usage without requiring the actual data to be fully exposed. This intermediary layer allows productivity to improve through public accessibility while data privacy is protected through automated, transparent access control mechanisms.
3Reliability
If cryptographic techniques are applied to protect AI models and datasets, then security is improved, but transparency and verification capability deteriorate
Solution Approach 1:
The patent implements feedback mechanisms through blockchain's immutable ledger and smart contract transparency. While cryptographic techniques protect the actual AI models and datasets, the blockchain provides continuous feedback about provenance, ownership, and access transactions. This feedback loop maintains transparency by recording all operations immutably on the blockchain, allowing verification of security measures without exposing the encrypted AI assets themselves.
Solution Approach 2:
The patent uses cryptographic hashing to create immutable copies of provenance and metadata information on the blockchain. These cryptographic copies (hashes) serve as verifiable representations of the original AI datasets and models without revealing the actual content. This allows transparency and verification capability to be maintained through these cryptographic copies while the original encrypted assets remain secure.
4Reliability
If multiple access control mechanisms are implemented for AI assets, then security is improved, but device complexity and operational difficulty increase
Solution Approach 1:
The patent implements self-service through automated smart contracts that handle access control, provenance tracking, and permission management without requiring manual intervention. The smart contracts automatically verify user credentials, enforce access policies, and record transactions on the blockchain. This automation reduces system complexity by eliminating the need for manual access control management while maintaining high security through cryptographic enforcement of policies.
Solution Approach 2:
The blockchain-based smart contract system serves multiple functions simultaneously: it provides access control, tracks provenance, manages ownership, and records usage statistics. This multi-functionality reduces overall system complexity by consolidating what would otherwise require separate complex systems into a single universal blockchain infrastructure that handles all security and tracking requirements through standardized smart contract mechanisms.
Data Source
AI summary
AI data and datasets that are represented as NFTs and carry all applicable data for a dataset's provenance, authenticity, and ownership. NFTs are used to validate datasets useful in AI training and can also be used to identify datasets that include faulty, biased, or otherwise erroneous data to improve predictive usefulness and reliability in decision making from the AI models.