Dynamic security generation AI system

Abstract

Traditional security systems relied on static authentication credentials, which increased the risk of data breaches and unauthorized access. Furthermore, once authentication credentials were leaked, they could be reused, leading to a decrease in security levels. Additionally, the lack of established technologies for dynamically generating authentication credentials in real time resulted in a lack of flexibility in responding to new cyberattacks. [Solution] This invention uses a generation AI to analyze user behavior data, device information, location information, and access time in real time, and generates a one-time-use authentication token. This token is dynamically generated using data from the device's sensors and network interface, and is invalidated after use. In addition, if abnormal access is detected, additional authentication measures are introduced to enhance security.

Description

Technical Field

[0001] The present invention relates to a dynamic security authentication system using Generative AI, and particularly belongs to a technology that analyzes user behavior data (e.g., operation patterns, access history), device information (e.g., terminal identification information, network information), and access environment information (e.g., geographical location, time zone, IP address) in real time, and dynamically generates and manages authentication information based on the analysis results.

[0002] Conventional security authentication systems rely on static authentication information (e.g., passwords, fixed tokens), and there is a problem that once these authentication information are leaked, the risk of unauthorized access and cyberattacks becomes extremely high. In addition, conventional AI technologies do not have sufficient ability to dynamically generate authentication information in real time, and the unpredictability against cyberattacks is limited.

[0003] In the present invention, these problems are solved by using a Generative AI model to analyze real-time user behavior data and environmental data, and dynamically generating an authentication token based on this data. The generated authentication token is limited to one-time use and reuse is prevented, so the security is significantly improved compared to static authentication methods.

[0004] Furthermore, the present invention applies ensemble learning technology to reduce the risk of false generation (hallucination) by Generative AI. By operating multiple Generative AI models in parallel and mutually verifying the outputs of each model, false generation can be prevented and the reliability of security authentication can be improved. This technology can be widely applied in fields such as financial institutions, medical institutions, e-commerce, and government agencies where security is strictly required.

[0005] Furthermore, since the system of the present invention is provided on a subscription basis, users can receive regular security updates and feature additions in real time. This format ensures that the system always maintains the latest security status and can respond quickly to newly emerging cyber threats.

[0006] The generated dynamic authentication token is provided to external systems in real time via an API (Application Programming Interface), and since API communication is protected by encryption technologies such as TLS (Transport Layer Security) and SSL (Secure Sockets Layer), the risk of authentication information leakage during communication is reduced.

[0007] Furthermore, the use of dynamic authentication tokens via APIs significantly reduces the risk of replay attacks and unauthorized access. Secure communication with external systems is ensured, enabling a highly secure environment that does not rely on traditional static authentication methods.

[0008] The technology of this invention can be applied to a variety of industries that prioritize security, such as finance, healthcare, e-commerce, and government agencies, and is particularly effective in situations where real-time security enhancement is required.

[0009] This invention also improves the user experience and streamlines security management by automating the dynamic authentication token generation and verification process. Automating dynamic token generation contributes to both enhanced security levels and reduced operational costs.

[0010] This invention provides a flexible and highly accurate authentication system that reduces security risks because dynamically generated authentication tokens interact with external systems and services in real time. [Background technology]

[0011] Traditional security authentication systems primarily rely on static authentication credentials, including passwords, PIN codes, and fixed tokens. However, a major problem with these static credentials is that once they are leaked, the risk of cyberattacks and unauthorized access becomes extremely high. If an attacker obtains these credentials, the authentication system has no means to prevent subsequent unauthorized access, severely compromising the system's security.

[0012] In response to this, AI technologies using anomaly detection and behavioral pattern analysis have been introduced in recent years. These AI systems analyze past data to detect abnormal behavior and signs of unauthorized access. However, these systems are mainly limited to reactive responses and have the limitation of not being able to handle the dynamic generation and management of security information in real time. In other words, existing AI technologies rely on anomaly detection based on past patterns and have not been able to provide sufficient unpredictability against new attack methods.

[0013] Furthermore, when transmitting static authentication information via an API (Application Programming Interface) in conjunction with external systems, the security of the communication cannot be fully guaranteed. In particular, if a fixed token or password is stolen, a replay attack using that information becomes possible, posing a risk of granting access to the external system. For this reason, even in communication via APIs, dynamic generation of authentication information and its invalidation each time were required.

[0014] Furthermore, while traditional security systems employed enhanced measures such as multi-factor authentication (MFA) and two-factor authentication (2FA), these were not foolproof. Although these methods provided additional authentication methods, they could not completely eliminate the risk of static information leakage and sometimes compromised the user experience. In particular, with two-factor authentication, delays in sending codes via SMS or email, and the risk of malicious third parties obtaining user authentication information have been pointed out.

[0015] Conventional security systems were unable to handle the real-time generation of dynamic authentication information, and authentication systems that relied on static authentication methods were particularly vulnerable in financial institutions, medical institutions, and government agencies where high levels of security are required. Furthermore, there were challenges in securely exchanging authentication information when communicating with external systems using APIs.

[0016] Against this backdrop, there was a need for technology that could provide authentication tokens that were dynamically generated in real time and could be used as disposable tokens. In addition, conventional generation AI technologies had a problem called hallucination (misgeneration), and if the generated authentication information was incorrect, it could cause a critical security risk.

[0017] To address the hallucination problem in conventional generative AI, ensemble learning techniques combining multiple models are considered effective. Ensemble learning significantly reduces the occurrence of hallucination by simultaneously running multiple generative AI models and mutually validating their outputs. This technique minimizes the security risks associated with conventional generative AI technologies.

[0018] Furthermore, even in API communication with external systems, dynamically generated authentication information is immediately invalidated after use, preventing replay attacks and the reuse of stolen authentication information. In addition, the API communication itself utilizes encryption technologies such as TLS and SSL, preventing the leakage of authentication information during transmission.

[0019] Based on this background technology, the main technical challenge of this invention is to provide a technology that enables the dynamic generation of security authentication information using generative AI and the reduction of hallucination risk through ensemble learning. Furthermore, this will enable the meeting of the advanced security requirements demanded by finance, healthcare, e-commerce, government agencies, and other sectors. [Prior art documents]

[0020] [Patent Document 1] Public Relations for Patent No. 2009532456

[0021] [Patent Document 2] Public Relations for Patent No. 2020054321 [Overview of the project] [Problems that the invention aims to solve]

[0022] Traditional security systems primarily rely on static authentication credentials (e.g., passwords, fixed tokens, PIN codes), which pose a significant problem: if this information is leaked, the risk of unauthorized access and cyberattacks increases dramatically. Furthermore, these static credentials, once obtained by an attacker, can be reused, posing a serious challenge as they easily weaken the security of the entire system.

[0023] Furthermore, conventional AI-based anomaly detection and pattern recognition technologies detected anomalies by analyzing past data. However, these technologies were unable to adequately address new attack methods and unpredictable security risks, and had limitations in generating dynamic security information in real time.

[0024] Furthermore, in systems utilizing generative AI technology, the risk of hallucination (misgeneration) during the generation process has been pointed out. If authentication tokens or security information are generated incorrectly, this could lead to cyberattacks or system failures, potentially resulting in a decrease in reliability.

[0025] In addition, when transmitting static authentication information via an API (Application Programming Interface) in communication with an external system, there is a risk of being targeted by cyberattacks such as replay attacks if the authentication information is stolen. If API communication is not sufficiently protected, there is a high risk that the fixed authentication information will be easily misused.

[0026] Conventional multi-factor authentication (MFA) and two-factor authentication (2FA) enhance security by combining multiple authentication means. However, these methods still rely on static elements, and in many cases, the user experience is impaired due to the complication of the authentication process itself. Also, the risk of leakage of static authentication information has not been completely avoided.

[0027] As described above, conventional security systems have multiple problems such as the vulnerability of static authentication information, the risk of hallucination by generative AI, the security in API communication with external systems, and the loss of user experience due to complex authentication processes. There has been a need for a technology that addresses these problems and provides a dynamic and highly reliable security system.

Means for Solving the Problems

[0028] In order to solve the above-described problems, the present invention provides a security system that analyzes user behavior data, device information, access environment information, etc. in real time using generative AI and dynamically generates and manages authentication tokens based thereon. With this system, since different dynamic authentication information is generated each time without depending on conventional static authentication information, the security risk is significantly reduced.

[0029] Furthermore, this invention reduces the risk of hallucination (misgeneration) caused by generative AI by applying ensemble learning technology. Specifically, by operating multiple generative AI models simultaneously and mutually verifying the output of each model, the probability of generating incorrect authentication information is kept extremely low. This technology ensures that dynamically generated authentication tokens are accurate, thereby improving the reliability of security authentication.

[0030] Furthermore, the generated authentication information is provided to external systems in real time via an API (Application Programming Interface). The generated tokens are limited to one-time use and are automatically invalidated after use, preventing replay attacks and misuse. API communication is secured by using encryption technologies such as TLS (Transport Layer Security) and SSL (Secure Sockets Layer).

[0031] The security system of the present invention generates the optimal authentication token on the spot based on the user's real-time behavioral and environmental data, resulting in high unpredictability and a dramatic improvement in security compared to conventional static authentication methods. Furthermore, if abnormal access or suspicious behavior is detected, it is possible to dynamically raise the security level by automatically applying additional authentication steps (e.g., multi-factor authentication) using ensemble learning.

[0032] This system also employs a subscription model, allowing users to receive regular security patches and new feature updates in real time. This ensures that the system maintains a security state that is always prepared to address the latest threats, while also reducing operational costs.

[0033] The dynamic security system based on the present invention is applicable to a variety of fields requiring high levels of security, such as finance, healthcare, e-commerce, and government agencies, and is expected to significantly reduce security risks and improve operational efficiency. [Effects of the Invention]

[0034] The benefit of this invention lies in the dramatic improvement in security reliability compared to conventional static security systems, achieved through the real-time dynamic generation of authentication tokens utilizing generative AI and ensemble learning. In particular, the following specific effects can be expected.

[0035] 1. Significant reduction in security risks This invention generates dynamic authentication tokens in real time based on user behavior data, device information, and access environment. These tokens are limited to one-time use and are immediately invalidated after use, resulting in a significantly lower risk of data breaches compared to traditional static passwords or fixed tokens. Furthermore, their non-reusability greatly reduces the risk of replay attacks and misuse.

[0036] 2. Reducing the risk of hallucination and improving reliability By applying ensemble learning, a mechanism is introduced in which multiple generative AI models mutually validate their outputs. This makes the probability of each model simultaneously generating incorrect information (hallucination) extremely low. This consensus-based approach by multiple AI models brings the risk of generating incorrect authentication information close to zero, dramatically improving the overall reliability of the security system.

[0037] 3. Real-time and dynamic security response The generation AI of this invention dynamically generates security information based on the user's real-time behavior and device information, making it more unpredictable than conventional static authentication systems. This makes it extremely difficult for attackers to guess or analyze authentication information in advance. Furthermore, if abnormal access or suspicious behavior is detected, the security level can be automatically increased and additional authentication steps can be applied, enabling a flexible and rapid response.

[0038] 4. Secure API integration with external systems In this invention, dynamically generated authentication tokens are provided to external systems in real time via an API (Application Programming Interface). This improves the security of communication with external systems because, instead of using fixed authentication information, a one-time token generated in real time is used. Furthermore, since the API communication itself uses encryption technologies such as TLS (Transport Layer Security) and SSL (Secure Sockets Layer), the risk of authentication information leakage is reduced.

[0039] 5. Reduce security operational costs The security system of this invention is provided on a subscription basis, allowing users to always maintain the latest security status. Regular security updates and new features are provided in real time, reducing the burden of manual maintenance and monitoring. Furthermore, increased automation in operation reduces personnel costs and system maintenance costs associated with security management.

[0040] 6. Applicability to diverse industrial fields This invention is highly effective in fields requiring a high level of security, such as financial institutions, medical institutions, e-commerce, and government agencies. Because dynamically generated authentication information adapts in real time to the specific security needs of each field, it has a wide range of applications. Its effectiveness is particularly evident in online transaction and confidential information management systems, where security is easily threatened.

[0041] 7. Improving the user experience The security system of the present invention dynamically adjusts the authentication process based on the user's access environment and behavior, providing seamless and rapid authentication under normal circumstances. On the other hand, if an anomaly is detected, an enhanced authentication procedure is automatically applied, making it possible to maintain a high level of security without compromising user convenience. conclusion

[0042] This invention realizes a real-time dynamic security system utilizing generative AI and ensemble learning, overcoming the limitations of conventional security systems and significantly improving reliability, flexibility, and efficiency. This makes it possible to achieve cost reduction and improved operational efficiency while maintaining a high level of security in various industrial sectors. [Brief explanation of the drawing]

[0043] [Figure 1] This is an overall system configuration diagram of the dynamic security generation AI system according to the present invention.

[0044] [Figure 2] This is a flowchart of the dynamic security generation AI system of the present invention. [Modes for carrying out the invention]

[0045] This invention relates to a security system that utilizes generative AI to generate and manage dynamic authentication information and security parameters in real time. This system analyzes real-time data such as user behavior data, device information, and access environment information, generates authentication tokens based on this data, and provides advanced security protection against cyberattacks and unauthorized access. Embodiments of this invention are described in detail below. 1. Generation of authentication tokens using a generation AI

[0046] The core technology of this invention lies in using generative AI to analyze user behavior data, device information, location information, and access environment, and dynamically generate authentication tokens based on this analysis. The generative AI generates tokens in real time when a user attempts to access the system. This enables unpredictable, dynamic authentication, unlike conventional security systems that rely on static passwords or fixed tokens.

[0047] This system uses an AI to analyze data obtained from the user's device (e.g., device ID, IP address, browser characteristics) and user behavior patterns (e.g., past access history, time of day, location), and generates the optimal authentication token in real time based on the results. The generated token is limited to one-time use and is immediately invalidated after use, thus preventing replay attacks and unauthorized reuse of tokens.

[0048] As an example of its application, in online banking systems at financial institutions, the risk of unauthorized access can be significantly reduced by using tokens generated based on past login history and current device information when users log in. Furthermore, in electronic medical record systems at healthcare institutions, the security of patient data can be ensured by using authentication tokens generated on the spot when healthcare professionals access the system. 2. Prevention of hallucination through ensemble learning

[0049] When a generative AI dynamically generates authentication tokens, there is a risk of hallucination (misgeneration). To address this problem, this invention introduces ensemble learning. Ensemble learning is a method in which multiple generative AI models are operated in parallel, and the tokens generated by each model are validated against each other.

[0050] Specifically, the system employs a mechanism where three or more generative AI models simultaneously generate tokens, and only if the results match are adopted as the final authentication token. This means that even if one model misgenerates, the other models will correct it, significantly reducing the risk of hallucination. This ensemble learning ensures that dynamically generated tokens are accurate and reliable, improving the reliability of security authentication.

[0051] Furthermore, if unusual access or suspicious behavior is detected, the generating AI can request a stricter authentication process than usual. For example, it can dynamically increase the security level by requiring additional multi-factor authentication (MFA) for the generated token. This functionality enables the system to achieve flexible and highly accurate security management. 3. Integration with external systems via API

[0052] The security system of the present invention can interact with external systems in real time via an API (Application Programming Interface). The authentication token generated by the generation AI is sent to an external authentication server or service via the API, and access is granted after its validity is verified.

[0053] In API communication, encryption technologies such as TLS (Transport Layer Security) and SSL (Secure Sockets Layer) are used to protect the communication, significantly reducing the risk of authentication information being leaked during transmission. Furthermore, since the generated token is limited to one-time use, even if the communication is intercepted, unauthorized reuse is prevented. 4. Anomaly detection and dynamic adjustment of security levels

[0054] This system not only authenticates users based on their normal access patterns, but can also automatically adjust security levels if unusual access or suspicious behavior is detected. For example, if a user attempts to access the system from an unusual device or location, the generated AI will determine this behavior as abnormal and request a more advanced authentication method (e.g., multi-factor authentication).

[0055] This anomaly detection function operates in real time and can flexibly respond to user behavior and access status, providing seamless authentication under normal circumstances while enabling rapid response when security risks increase. This allows for maintaining the security of the entire system without compromising the user experience. [Examples] Example 1: Dynamic Authentication System in a Financial Institution

[0056] In online banking and transaction systems at financial institutions, security is a top priority, and advanced authentication systems are required to prevent unauthorized access. This embodiment describes an example of online banking using the dynamic security generation AI system of the present invention.

[0057] When a user attempts to log in to online banking, the generating AI first collects the user's device information (e.g., device ID, IP address, browser information) in real time. In addition, it analyzes the user's behavioral patterns based on their past access history (e.g., typical login times and geographical location), and generates a one-time-use authentication token based on the results.

[0058] The generated authentication token is sent to the bank's authentication server via API. The server verifies the token's validity and confirms that the user is accessing from the correct device. This token is invalidated immediately after use, preventing replay attacks. If unusual behavior (e.g., access from an unusual device or IP address) is detected, the generating AI can automatically raise the security level and require additional multi-factor authentication (e.g., SMS authentication, email authentication).

[0059] By implementing this dynamic authentication system, banks can provide flexible and advanced security that adapts to user access in real time, without relying on traditional passwords or static tokens. Furthermore, because authentication tokens are limited to one-time use, misuse and reuse of tokens can be completely prevented. Example 2: Patient data protection in healthcare facilities

[0060] In healthcare institutions, protecting patients' personal information and medical records is of paramount importance, and ensuring security is a top priority. This embodiment describes the authentication process when healthcare professionals access electronic health records (EHR / EMR) using the dynamic security generation AI system of the present invention.

[0061] When a healthcare professional attempts to access an electronic medical record, the generating AI creates an authentication token based on the professional's device information (e.g., hospital terminal, personal tablet), location information (e.g., inside or outside the hospital), and working hours. This token is dynamically generated in real time according to the healthcare professional's access permissions, and grants them permission to access the medical record.

[0062] The generated authentication token is verified through the healthcare institution's internal system to confirm that the employee has legitimate access rights. This token is immediately invalidated after a single use, so even if a third party illegally obtains the authentication information, they cannot reuse it. Furthermore, if unusual access (e.g., access outside of normal working hours or access from an unusual device) is detected, the system automatically raises the security level and requests additional authentication processes (e.g., multi-factor authentication).

[0063] This system enables healthcare institutions to prevent unauthorized access to patient data through real-time, dynamic authentication, and to securely manage highly confidential medical information. Furthermore, it allows healthcare professionals to perform their duties more efficiently by minimizing cumbersome authentication processes while enhancing the overall security of the system. Example 3: Preventing fraudulent transactions on an e-commerce platform

[0064] E-commerce platforms require robust authentication systems to protect customer account and payment information. This embodiment describes a customer authentication process utilizing the dynamic security generation AI system of the present invention.

[0065] When a customer logs into an e-commerce site and attempts to purchase a product, the generating AI creates a dynamic authentication token in real time based on the customer's device information (e.g., smartphone, PC), past purchase history, IP address, and geographical location. This token is for one-time use only and is invalidated once the payment process is complete.

[0066] The generated authentication token is sent to the e-commerce system via the API, verified by the system, and then the transaction is approved. If the system detects behavior that deviates from the customer's usual behavior patterns (e.g., login from a new device, access from an unusual region), it will request additional multi-factor authentication to ensure the security of the transaction.

[0067] By utilizing this system, e-commerce platforms can effectively prevent fraudulent transactions without compromising the customer experience. Dynamically generated authentication tokens are far more secure than traditional static passwords, particularly enhancing security in the payment process. Example 4: Protecting confidential information in government agencies

[0068] In government agencies, protecting confidential information is of paramount importance, and it is essential to thoroughly prevent unauthorized access from external sources. This embodiment describes an authentication process in a government agency system utilizing the dynamic security generation AI system of the present invention.

[0069] When a government employee attempts to access classified information, the generating AI creates a dynamic authentication token based on data such as the employee's access device, location, and time of day. The generated token is optimized according to the employee's position and access privileges to ensure system security.

[0070] This authentication token is for one-time use only and is sent to the sensitive information system via the API. If an unusual access pattern or access from an unusual device is detected, the generating AI will request additional multi-factor authentication to enhance the overall system security.

[0071] This system enables government agencies to prevent unauthorized access to confidential information and provide high-level security in real time, without relying on traditional static authentication credentials. Furthermore, dynamic authentication using generated AI minimizes security risks while maintaining employee productivity. [Industrial applicability]

[0072] The dynamic security generation AI system according to the present invention is a system that significantly reduces security risks and achieves advanced security protection by dynamically generating authentication tokens using generation AI and ensemble learning technology. This technology can be widely applied in multiple industrial fields, and is particularly expected to be used in the following fields. 1. Financial industry

[0073] The financial industry constantly demands high levels of security in areas such as online banking, securities trading, and digital payment systems. By introducing the system of the present invention into the financial sector, it is possible to prevent unauthorized access and cyberattacks by utilizing authentication tokens that are dynamically generated in real time based on user access behavior and device information.

[0074] Specifically, when a customer logs into online banking, the generating AI analyzes past access history and device information to generate a new authentication token in real time, making it difficult for attackers to access the system using passwords or tokens they have previously stolen. Furthermore, if unusual access or suspicious activity is detected, the security level is automatically raised through ensemble learning, and additional authentication processes are introduced, providing a higher level of security.

[0075] By implementing this system, financial institutions can significantly reduce the risk of unauthorized access, enhance customer trust, and expect to reduce operational costs associated with maintaining security. 2. Healthcare Industry

[0076] In the medical industry, protecting patients' personal information and medical records is extremely important, and a high level of security is required. By introducing the system of the present invention into the medical field, the risk of unauthorized access and data leakage can be reduced when medical professionals access electronic medical record systems by using dynamically generated authentication tokens.

[0077] Traditionally, healthcare institutions have used passwords or fixed tokens for access control. However, these static authentication credentials are vulnerable to leakage and cyberattacks. This invention enables a more secure authentication process that does not rely on static authentication credentials by utilizing authentication tokens generated in real time based on the device information, working hours, and access environment of healthcare professionals.

[0078] Furthermore, the generated tokens are limited to one-time use and are immediately invalidated after use, preventing the reuse of authentication information. This effectively prevents unauthorized access to patient data and also contributes to increased efficiency for healthcare professionals by eliminating the need for cumbersome authentication processes. 3. E-commerce and online services industry

[0079] In e-commerce and online services, protecting customer account and payment information is extremely important, and security vulnerabilities are highly likely to lead to direct damage. By introducing the system of the present invention, when customers log in or make payments, they can use dynamically generated authentication tokens based on their past purchase history and device information, thereby preventing fraudulent transactions and account hijacking.

[0080] In particular, traditional authentication systems using passwords or fixed tokens posed a risk of misuse if this information was leaked. However, the present invention utilizes dynamic tokens that are limited to one-time use, making it impossible to reuse leaked tokens. This effectively prevents phishing attacks and fraudulent transactions.

[0081] Furthermore, if unusual purchase patterns or unauthorized access are detected, the generating AI can automatically request additional authentication processes, thereby raising the security level. This improves the security of e-commerce platforms and online services without compromising the customer experience. 4. Government agencies and public services

[0082] Government agencies and public services require the highest level of security to protect citizens' personal and confidential information. By introducing the system of the present invention, it is possible to prevent unauthorized external access by using dynamically generated authentication tokens in the authentication of employees and citizens accessing government agency systems.

[0083] For example, when citizens access tax systems or resident services online, the AI ​​generates an authentication token based on their access environment and device information, and sends it to the system via an API. This provides far greater security than traditional authentication methods using passwords or fixed tokens. Furthermore, when employees access confidential information, if behavior deviating from normal access patterns is detected, additional multi-factor authentication is required, thus preventing unauthorized access to confidential information.

[0084] Thus, the system of the present invention can minimize security vulnerabilities while promoting the digitalization of government agencies and public services by performing dynamic authentication in real time. 5. The telecommunications industry and cloud services

[0085] In the telecommunications and cloud service industries, vast amounts of data are exchanged over networks, making security crucial. The system of the present invention can be used in authentication processes for accessing cloud services and communication networks, and by using authentication tokens generated in real time based on the user's device information and access status, network security can be enhanced.

[0086] In particular, when accessing cloud services, numerous users connect from various devices, making traditional methods that rely on fixed authentication information highly risky. This invention provides more flexible and robust security because a different authentication token is generated in real time for each user. Furthermore, if abnormal access is detected, an additional authentication process is dynamically applied, enhancing the security of the entire communication infrastructure. conclusion

[0087] The dynamic security generation AI system of the present invention can be applied to a variety of industries with high security risks, including finance, healthcare, e-commerce, government agencies, and telecommunications. By dynamically generating authentication tokens in real time using generation AI and ensemble learning technology, security is dramatically improved compared to conventional static authentication methods, enabling data protection and security maintenance in each industry. Furthermore, because it can provide advanced security without compromising the user experience, it also contributes to operational efficiency and cost reduction.

Claims

1. A security system that uses generation AI to generate dynamic authentication information in real time based on user behavior data, device information, location information, and access time, and provides this authentication information to external systems via API as a one-time-use token. The system is characterized by the fact that the generated authentication information is invalidated after use and cannot be reused.

2. The security system according to claim 1 is characterized in that the generated authentication token is adjusted based on the user's physical environment (e.g., device information, IP address, location information, time zone).

3. The security system according to claim 1 or 2 is characterized in that the generated authentication token is invalidated after specified conditions (e.g., time, location, usage status) are met.

4. The security system according to any one of claims 1 to 3 is characterized in that, when abnormal access behavior is detected, the generating AI dynamically generates an additional authentication token and requests multi-factor authentication from the user.

5. A security system characterized by using QR codes or NFC tags to dynamically verify authentication information in real time, and denying access if authentication fails.

6. A security system characterized by having a function that detects abnormal communication or signal delays in dynamically generated authentication tokens and invalidates the authentication tokens in order to prevent relay attacks.

7. The security system according to claim 1 is characterized in that the authentication token provided to an external system in real time via an API is encrypted using TLS or SSL, thereby ensuring the security of the communication.

8. The security system according to claim 5 is characterized by having a function to generate dynamic authentication information for QR codes and NFC tags, and to prevent unauthorized scanning and tag tampering.

9. The security system according to claim 5 or 8 is characterized by having a function to deny access if authentication of the QR code or NFC tag fails.

10. This security system is applied to cloud infrastructure and is characterized by its ability to generate dynamic authentication information based on the user's connection status using AI, thereby managing secure access to cloud applications.

11. The security system described in claim 10 is provided on a subscription basis, and the user receives security updates and new features in real time on a regular basis, thereby maintaining the latest security status at all times.

12. A security system characterized by the use of a generated authentication token in combination with the user's biometric authentication information (fingerprint, facial recognition, etc.).

13. The security system according to claim 1 is characterized in that the generated authentication information is adjusted in real time based on the user's behavior patterns and past access history.

14. The security system according to claim 11 is characterized by receiving security updates periodically on a subscription basis.

15. A security system characterized by its relay attack prevention function and its application to automobile door lock systems.

16. A security system for an automobile ignition system characterized by invalidating the authentication token when a relay attack is detected.

17. The security system according to claim 1 is characterized in that, when accessed from different devices, tokens are dynamically generated based on each device, and seamless authentication is performed.

18. A security system that uses generation AI to generate dynamic authentication information in real time based on user behavior data, device information, location information, and access time, and provides this authentication information to external systems via API as a one-time-use token. The system is characterized by the fact that the generated authentication information is invalidated after use and cannot be reused.

19. The security system according to claim 18 is characterized by applying ensemble learning technology, operating multiple generative AI models in parallel, comparing and verifying the authentication information generated by each model to prevent hallucination, and ultimately adopting the most reliable authentication information.

20. The security system according to claim 19 is characterized in that, when abnormal access or behavior is detected, the security level is dynamically increased using ensemble learning and additional multi-factor authentication is requested.

21. The security system according to claim 19 is characterized by having a function that compares the outputs of multiple models obtained by ensemble learning, generates a token only when the authentication information matches, and detects an anomaly if a mismatch occurs.

22. The security system according to any one of claims 18 to 21 is characterized in that the generated authentication token is adjusted based on the user's physical environment (device information, IP address, location information, time zone).

Images