Persona authentication system

Abstract

To solve the problems of current authentication systems that face limitations in vulnerabilities and usability due to development of AI or generative AI, despite advancements in multi-factor authentication, facial recognition or voice recognition.SOLUTION: Problems in an authentication system can be solved by combining multiple authentication functions regarding conversational habits, use of memories, and daily behavioral patterns. Specifically, conversational authentication, memory-based authentication, behavioral authentication, context authentication, and other various biometric authentication methods are integrated using AI and generative AI, to provide various authentication systems, such as stepwise authentication, adaptive authentication or continuous authentication, thereby enhancing security and significantly increasing reliability. A learning function can cope with the change of users, such as ages, thereby improving overall usability of authentication processes and increasing the possibility of development to various services.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an authentication system, a management device, and an authentication method that can efficiently authenticate that a person performing various procedures is a legitimate user (hereinafter referred to as "personal authentication") when the user performs the procedures. [Background technology]

[0002] Conventional authentication methods for logging into systems include password authentication, biometric authentication using biometric information such as fingerprints or faces, token authentication using physical devices such as mobile terminals, and multi-factor authentication that combines these methods.

[0003] However, there are already established techniques to break these authentication techniques, resulting in problems with account fraud. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7408882 [Non-patent literature]

[0004] [Non-Patent Document 1] "Can a facial recognition system be bypassed with a photo? A detailed explanation of the methods and countermeasures" SECURITY MEDIA (July 21, 2023)

[0005] [Non-patent document 2] "How Multi-Factor Authentication is Broken: Five Major Attack Methods and Defenses," ITmedia Inc. (April 16, 2024) Summary of the Invention [Problem to be solved by the invention]

[0006] In the aforementioned facial recognition, technology is used to break authentication by "impersonating" someone by "using photographs or videos" or "using 3D masks or mannequins," and to prevent impersonation, "a system requiring two-factor authentication" is said to be required.

[0007] However, as mentioned above, there are already technologies that can break multi-factor authentication by "spoofing" using "SIM swapping" and "man-in-the-middle attacks," so there is still room for improvement in "identity verification" within "authentication," and a more reliable and efficient system for "identity verification" and "authentication" is needed.

[0008] The present invention has been made to solve the problems of the above-mentioned conventional technology, and aims to provide an authentication system, management device, and authentication method that can reliably and efficiently perform authentication when a user performs various procedures. [Means for solving the problem]

[0009] In order to solve the above problems, the present invention provides an authentication system having an authentication information management means that uses AI and generation AI technology to manage multiple pieces of authentication information for authenticating users, and is equipped with a usage type determination means that determines a predetermined usage type, an authentication type selection means that selects from multiple authentication types an authentication type that corresponds to the usage type determined by the usage type determination means, an authentication processing means that performs authentication processing of the user based on the authentication type selected by the authentication type selection means, and a notification means that notifies the user of the authentication result by the authentication processing means.

[0010] In the present invention, in the above-mentioned invention, the authentication processing means performs authentication processing including at least dynamic biometric authentication using a moving image including a face portion of the user and the voice of the user.

[0011] Furthermore, in the present invention, in the above invention, the dynamic biometric authentication performs authentication processing including face authentication processing based on a face image of the user contained in the moving image, and voice authentication processing based on voice information of the user contained in the moving image.

[0012] The above-mentioned problems of the present invention are solved by the "authentication functions (1) to (6)" and "authentication methods (7) to (9)" that use the following AI and generation AI.

[0013] (1) Conversational authentication: Recognizes the user's conversational habits and characteristics and authenticates them using text and voice data.

[0014] (2) Memory authentication: Patterns are extracted from the user's past conversation records and input history and used in the authentication process.

[0015] (3) Behavioral authentication: Authentication is performed using subtle behavioral characteristics, such as the way a smartphone is held or typing rhythm.

[0016] (4) Contextual authentication: Optimizes the authentication process based on context, such as location, ambient sounds, and time of day.

[0017] (5) Diversification of biometric authentication: Combining multiple biometric authentication technologies, such as facial recognition, fingerprint, and iris recognition.

[0018] (6) Learning function: The system learns changes in the user's facial expressions and voice quality, and maintains authentication accuracy according to changes in age and emotions.

[0019] (7) Step-by-step authentication method: A method that applies different authentication functions in stages based on the authentication result and risk assessment. Low-risk operations require basic authentication, and high-risk operations incorporate additional authentication steps.

[0020] (8) The adaptive authentication method: Dynamically selects the necessary authentication method based on context information such as the user's behavior, environment, and time of day, and automatically adjusts the authentication process.

[0021] (9) Continuous authentication method: A method in which the user's behavior is monitored continuously or intermittently while logged in to the system, and re-authentication is required if abnormal behavior is detected. This addresses security risks that occur during a session. [Effects of the Invention]

[0022] In this invention, by adding "internal characteristics" such as "conversational habits and features" and "user memory" to the conventional "external characteristics" such as the user's face and voice, as well as "subtle external characteristics" such as "behavioral characteristics" and "combination of multiple biometric authentication technologies," and even "environmental characteristics," and authenticating using AI and generative AI technology, it is possible to dramatically improve the "accuracy of identity verification" and the "difficulty of impersonation, etc."

[0023] By using AI and generative AI technology to continuously learn authentication information, it becomes possible to authenticate users regardless of changes in their personalities due to aging, illness, etc.

[0024] Diversifying authentication methods enables flexible and convenient authentication that is tailored to the "user situation" and "security policies for each system."

[0025] Devices equipped with image recognition functions (camera, etc.), voice input / output functions (microphone / speaker), and communication functions (Wi-Fi, etc.) can be used for authentication alone, eliminating the need for a separate device as in multi-factor authentication, and eliminating the risk of use due to malfunction or loss of the device. [Brief explanation of the drawings]

[0026] [Figure 1] Flowchart for the multi-level authentication method [Figure 2] Flowchart for adaptive authentication method [Figure 3] Flowchart for sequential authentication method DETAILED DESCRIPTION OF THE INVENTION

[0027] The present invention can be embodied in three forms (1) to (3) described in the following paragraphs.

[0028] (1) The system to be used is equipped with the authentication function, and authentication is performed when the system is used.

[0029] (2) The authentication function is systemized as SaaS, etc., and authentication is performed using the authentication system when authenticating each system.

[0030] (3) It is introduced as an edge computing tool on the terminal to be used, and authentication is performed when the terminal is used, and authentication information is sent to each system. [Industrial Applicability]

[0031] By using the "authentication function" as a "key" for "identity verification," and by adding functions such as "authentication history" and "emotion analysis using AI and generative AI," and by linking with existing services, it is possible to expand into the following services. 1. Authentication (1) Login authentication services for various IoT systems On-premise system IaaS, PaaS, and SaaS services ○SEAM (2) Security floor entry / exit authentication, log management, and health management services ○Personal authentication ○Emotion recognition Physical and mental health management Recognition and learning of user facial expressions, voice quality, etc. ○ Integration with existing services and devices · Temperature measurement (COVID-19 prevention measures) - Linkage with attendance management systems, payroll systems, etc. (3) Authentication services for house keys, etc. ○Locking and unlocking the front door ○Locking and unlocking windows, etc. ○ Integration with existing services and devices Remote locking Intercom Security cameras · Home security camera (for pets, etc.) *Combined with existing locking methods (4) Authentication services for safe keys, etc. ○Locking and unlocking safes, etc. *Combined with existing locking methods Integration with IoT services and devices Remote lock confirmation, etc. -Notification device when unlocked (5) School and cram school attendance history and health management services ○Installed in classrooms and other locations on campus ○Installed on school routes such as stations and bus stops (6) Customer identification and emotion recognition services ○ Restaurants and hospitality businesses *Terminals installed in elevators, etc. → Customer service preparation possible before the customer arrives (preparation for visit, order history, etc.) (7) Care record management services for home care, in-room care, etc. ○ Record and manage the "start and end" of assistance and "contents of assistance" * Caregiver attendance management, payroll calculation, care reporting to municipalities, care Including cooperation with the insurance premium application system ○ Recording and service of "physical condition and emotional analysis" of users during care Sharing with the provider ○ Support for caregivers in the event of assistance methods or malfunctions, and service provision Support for contacting the responsible person (8) Authentication services for vehicle keys, etc. ○Authentication for locking and unlocking vehicle doors, starting the engine, etc. ○ Notification function by linking with dashcams, security cameras, etc. ○ Attendance management systems for automobile driving (trucks, etc.) Linking with payroll systems, etc. *2024 issue (limits on overtime work in transportation and other fields) Thorough attendance management in 2. Crime prevention (1) "Personal authentication" and "AI-based fraud prevention" when using bank and postal savings ATMs ○When using ATM *Preventing fraud damage by determining whether or not a mobile phone is used *Reduce fraud damage by comparing usage history etc. using AI ○When using a smartphone, etc. *Reduce fraud damage by comparing usage history etc. using AI (2) "Personal authentication" and "AI" when using credit cards, electronic money, etc. "Fraud prevention through ○ In-store payment (unmanned / manned register) ○Smartphone payments, etc. * AI reduces fraud damage by estimating past usage history, etc.

Claims

1. An authentication system using AI and generation AI having authentication information management means for managing multiple pieces of authentication information for authenticating users, comprising: usage type determination means for determining a predetermined usage type; authentication type selection means for selecting from multiple authentication types an authentication type corresponding to the usage type determined by the usage type determination means; authentication processing means for performing authentication processing of the user based on the authentication type selected by the authentication type selection means; notification means for notifying the authentication result of the user by the authentication processing means; and a function for learning information at the time of authentication.

2. The authentication system according to claim 1, wherein the authentication processing means includes at least dynamic biometric authentication using the user's personal characteristics using AI and generated AI, as well as a moving image including the user's face and voice, and authentication using environmental information, and selects the most appropriate authentication method at the time of authentication processing from a plurality of authentication methods to perform the authentication processing.

3. 3. The authentication system according to claim 2, wherein the dynamic biometric authentication has a conversation authentication function, a memory authentication function, a behavioral authentication function, and a context authentication function that utilize AI and generative AI based on the user's authentication input information contained in the video, voice, and text input, an integrated authentication function of various biometric authentication technologies, and a learning function of authentication information.

4. The authentication system of claim 2, wherein the plurality of authentication methods are a combination of conventional authentication technology and the authentication function of claim 3, and by utilizing AI and generated AI, combines optimal authentication functions and authentication procedures taking into account the situation at the time of authentication.

5. 4. The authentication system of claim 3, wherein the conversational authentication function is an authentication function that performs authentication based on the user's writing style and conversational phrasing, and this function identifies user-specific expression patterns from text or voice input and utilizes these in the authentication process.

6. 4. The authentication system of claim 3, wherein the memory-based authentication function utilizes text or voice content previously input or generated by the user to perform authentication, and the function learns from the user's past data and uses past input patterns as a comparison standard during the authentication process.

7. The authentication system according to claim 3 , wherein the behavioral authentication function analyzes the user's behavioral patterns (such as how to hold a smartphone, typing rhythm, walking pattern, etc.) and uses the results for authentication.

8. The authentication system of claim 3 , wherein the contextual authentication function incorporates the user's current situation and environmental information (such as location, ambient sounds, time of day, etc.) into the authentication process.

9. The authentication system according to claim 3 , wherein the integrated authentication function of the various biometric authentication technologies utilizes a combination of a plurality of biometric authentication technologies, such as face authentication, voiceprint authentication, fingerprint authentication, iris authentication, and palm vein authentication, for authentication.

10. The authentication system according to claim 3, wherein the learning function learns changes in the user's facial expression and voice quality during authentication and reflects the changes in the authentication process, thereby maintaining authentication accuracy while adapting to natural changes in the user's behavior.

11. 5. The authentication system of claim 4, wherein the tiered authentication scheme applies different authentication functions in stages based on authentication results and risk assessment, requiring basic authentication for low-risk operations and incorporating additional authentication steps for high-risk operations.

12. The authentication system according to claim 4 , wherein the adaptive authentication method dynamically selects a required authentication method and automatically adjusts the authentication process based on context information such as user behavior, environment, and time of day.

13. 5. The authentication system according to claim 4, wherein the continuous authentication method continuously or intermittently monitors the user's behavior while the user is logged in to the system, and requests re-authentication if abnormal behavior is detected, thereby addressing security risks that occur during a session.

Images