AI for military defense purposes

Abstract

Traditional military defense systems rely on static authentication credentials, which means that a single leak increases the risk of cyberattacks and unauthorized access. Furthermore, their limited anomaly detection and dynamic authentication capabilities make rapid response difficult. [Solution] This invention is a system that uses generative AI to analyze user behavior data and location information, and dynamically generates one-time tokens for authentication in military defense systems. Specialized for attack drones, defensive drones, and missile defense systems, an anomaly detection module detects unauthorized access in real time and raises the security level as needed. This provides flexible and responsive security measures for each defense system.

Description

Technical Field

[0001] The present invention relates to a dynamic security generation system using Generative AI, and particularly belongs to a technology for strengthening the cyber security of military defense systems. The present invention provides a new security method in a military defense system that provides dynamically generated authentication information and communication protection to prevent unauthorized operations due to external cyber attacks and hacking.

[0002] The dynamic security generation system based on the present invention enables an authentication process specialized for military defense systems and dynamic security responses, realizing advanced security protection to prevent leakage of military confidential information, unauthorized access to the system, and invalidation. In addition, the responsiveness to threats can be improved through security tokens and communication encryption dynamically generated by Generative AI.

[0003] The present invention overcomes the limitations of existing static security methods and provides a dynamic security generation system specialized for military defense systems, which does not depend on fixed authentication information, reduces security risks, and improves the reliability and flexibility of the system.

Background Art

[0004] Security in conventional military defense systems mainly relied on static authentication information (e.g., passwords, fixed tokens, encryption keys). These static authentication information had a high risk of leakage and hacking, and once an attacker obtained the authentication information, they could easily perform unauthorized access and system operations. In addition, there was a risk that information requiring confidentiality in military systems would leak outside, and the vulnerability in security management had become a major issue.

[0005] Furthermore, traditional security methods lacked the flexibility to respond to new cyberattack techniques and unpredictable threats because authentication credentials were fixed. In particular, military defense systems required the ability to respond to threats in real time, but systems using fixed authentication credentials had limitations in their ability to quickly evade enemy attacks and strengthen system defenses.

[0006] In addition, while AI-powered security technologies have been introduced in recent years to prevent unauthorized access through anomaly detection and pattern recognition, these approaches are primarily based on historical data and have not been able to adequately address new threats occurring in real time or the dynamic generation of authentication information in response to changing circumstances. Furthermore, existing AI-based security technologies have limitations in providing sufficient immediate response to adversarial intrusions because authentication information is in a fixed format.

[0007] Even in systems utilizing generative AI, there is a risk of hallucination during the generation process. If incorrect authentication information is generated, the system's defensive capabilities may be compromised. In the military field in particular, the risks posed by hallucination in generative AI are extremely high, and reliability is paramount. Furthermore, if the generated authentication information is inaccurate, the likelihood of enemy tampering with defense systems or information leaks increases.

[0008] Therefore, military defense systems require technology to dynamically and in real time generate authentication information and reduce the risk of unauthorized access. Furthermore, there is a need for a dynamic security generation system that reduces the risk of misgeneration by the generating AI and enables real-time security enhancements tailored to each military defense system. This will provide the latest defenses against cyberattacks and unauthorized external system manipulation of military systems, thereby enhancing the safety and reliability of the system. [Prior art documents] [Patent Documents]

[0009] [Patent Document 1] Public Relations for Patent No. 5375310 [Patent Document 2] Public Relations for Patent No. 6892456 [Overview of the Initiative] [Problems that the invention aims to solve]

[0010] In military defense systems, preventing cyberattacks and unauthorized external access is crucial. However, traditional security systems primarily rely on static authentication information (passwords and fixed tokens), which, once leaked, compromises system confidentiality and weakens defense capabilities. Furthermore, systems using static authentication information have limitations in responding to the diversification of threats and real-time attack methods, making it difficult to provide adequate defense against the latest cyberattacks.

[0011] Furthermore, although AI-based anomaly detection technology has been introduced, its reliance on historical data means it cannot immediately respond to new threats or unpredictable situations. In the military field, where high reliability is required, the risk of hallucination (misrepresentation) by generating AI is extremely high. If misrepresented authentication information is adopted, there is a risk that defenses against external tampering will not function adequately. Also, if authentication information is inaccurate, the risk of defense systems against hostile forces being disabled increases.

[0012] As described above, military defense systems have required security technology that dynamically generates authentication information using generative AI, without relying on static authentication, thereby reducing the risk of unauthorized access and manipulation. The present invention aims to solve these problems and provide real-time and highly reliable security measures. [Means for solving the problem]

[0013] This invention provides a system that dynamically generates and manages one-time-use authentication tokens by analyzing real-time data such as user behavior data, device information, location information, and access time, in order to dynamically enhance the security of military defense systems using generative AI. Since these authentication tokens are immediately invalidated after use, the risk of replay attacks and unauthorized reuse is reduced, and a high level of security that does not rely on conventional fixed authentication information is achieved.

[0014] Furthermore, this invention employs ensemble learning technology to reduce the risk of misgeneration by generative AI. By operating multiple generative AI models simultaneously and mutually verifying the output of each model, the probability of misgeneration is kept extremely low, providing accurate and reliable authentication information. This minimizes the risk of misgeneration in the authentication process, enabling immediate response to real-time threats.

[0015] Furthermore, the dynamic security generation system of the present invention can cooperate with external authentication servers and other systems via APIs using encryption technologies such as TLS (Transport Layer Security) and SSL (Secure Sockets Layer). This reduces the risk of authentication information leakage in the communication path and enables dynamic and responsive security responses in each military defense system.

[0016] Furthermore, the system of the present invention is equipped with a real-time anomaly detection function in military defense systems, and can dynamically raise the security level when abnormal access or suspicious behavior is detected. Specifically, it can enhance security strength and immediately implement defensive measures by requiring additional multi-factor authentication (MFA).

[0017] The system of this invention prevents the leakage and unauthorized access of classified military information, while achieving flexible and reliable defensive responses by combining generated AI and encrypted communication, providing advanced security that does not rely on conventional static authentication information. As a result, security risks to military defense systems are significantly reduced, and the system's safety and responsiveness are dramatically improved. [Effects of the Invention]

[0018] The dynamic security generation AI system according to the present invention eliminates the use of static authentication information in military defense systems and generates and manages one-time authentication tokens based on the generation AI in real time, thereby significantly reducing the risk of unauthorized access and cyberattacks compared to conventional security methods. This ensures the reliability and confidentiality of the defense system and allows for flexible responses to the latest threats.

[0019] By applying ensemble learning technology to generative AI, the risk of hallucination (misgenerative errors) seen in conventional generative AI is significantly reduced. Because the system of this invention has multiple generative models simultaneously verifying their outputs, the probability of each model simultaneously generating a misgeneration is extremely low, providing highly reliable authentication information in real time. This functionality enables immediate response to unpredictable new threats and unauthorized access, dramatically improving system security.

[0020] Furthermore, the system of the present invention utilizes encryption technologies such as TLS (Transport Layer Security) and SSL (Secure Sockets Layer) for API communication, preventing the leakage of authentication information during communication with external systems while enabling secure, real-time information exchange. This encrypted communication reduces the risk of cyberattacks and man-in-the-middle attacks, thereby strengthening information security in military defense systems.

[0021] In addition, the dynamic security generation system of the present invention has a real-time anomaly detection function. When suspicious behavior or abnormal access is detected, the security level can be immediately raised. With this function, while providing fast and seamless authentication during normal access, when a security risk occurs, enhanced defensive measures are automatically applied, ensuring the security of the system.

[0022] The present invention not only greatly improves the security strength of military defense systems, but also enhances the flexibility of system operation by combining generative AI and dynamic authentication information generation. As a result, military defense systems can always maintain a security state capable of responding to the latest cyber threats, reduce management costs compared to conventional static authentication systems, and provide a highly reliable defense mechanism over the long term.

Brief Description of the Drawings

[0023] [Figure 1] It is an overall system configuration diagram of the dynamic security generation AI system according to the present invention. [Figure 2] It is a flowchart in the dynamic security generation process of the present invention.

Modes for Carrying Out the Invention

[0024] The present invention relates to a security system for enhancing the security of military defense systems by utilizing generative AI to generate dynamic authentication information in real time. Hereinafter, embodiments of the present invention will be described in detail for each of the main components of the generative AI model, anomaly detection module, and encrypted communication interface.

[0025] First, the generative AI model, which is the central element of this invention, analyzes various real-time data related to the military defense system, such as user behavior data, device information, location information, and access time, and generates a one-time authentication token tailored to the situation. The generative AI model generates the token in real time when accessing the system and invalidates it immediately after use. This dynamic generation significantly reduces the risk of cyberattacks and replay attacks, unlike conventional static authentication information.

[0026] Furthermore, the generative AI model utilizes ensemble learning technology, which combines multiple generative algorithms, achieving high reliability to suppress the risk of hallucination. Specifically, multiple generative AI models simultaneously generate authentication tokens, and the integrity of the tokens generated by each model is mutually verified. This mechanism ensures that even if a misgeneration occurs in one of the models, the other models correct it, providing accurate and reliable authentication information.

[0027] The anomaly detection module is responsible for detecting suspicious access and abnormal behavior that deviates from normal access patterns within military defense systems in real time. Based on data analyzed by a generative AI model, this module automatically identifies unexpected behavior and suspicious actions, and has the ability to raise the security level as needed. Specifically, if an anomaly is detected, it can request additional multi-factor authentication (MFA) to prevent unauthorized access to the system. This anomaly detection function allows for fast and seamless authentication for normal access, while immediately applying enhanced security measures when the risk of unauthorized access is high.

[0028] The system of the present invention is equipped with an encrypted communication interface for securely coordinating with external authentication servers and other systems. This interface uses encryption technologies such as TLS (Transport Layer Security) and SSL (Secure Sockets Layer) for API communication, mitigating the risk of authentication information leakage and man-in-the-middle attacks during communication. Authentication tokens dynamically generated by the generation AI model are sent to external systems via this encrypted communication and are immediately invalidated after transmission, thus preventing unauthorized reuse.

[0029] Furthermore, since the system of the present invention is provided on a subscription basis, users can receive regular security patch applications and new features in real time. This format ensures that the system is always ready to respond to the latest cyber threats, resulting in reduced operational costs and more efficient management. The dynamic security generation system of the present invention, specifically designed for military defense systems, provides a highly responsive and reliable defense system in today's world of increasingly diverse threats and sophisticated attacks. [Examples]

[0030] The following embodiments describe a specific method for protecting a military defense system from cyberattacks and unauthorized access by applying the dynamic security generation AI system of the present invention to the system. The present invention features dynamic generation of authentication tokens using generation AI and real-time anomaly detection capabilities, achieving advanced security protection in the military field. Example 1: Application to military communication systems

[0031] The dynamic security generation system of the present invention is used in military communication systems to dynamically authenticate each communication in real time. The generation AI model analyzes user behavior data, device information, location information, etc., in real time and generates a one-time authentication token based on that information. The token generated before the start of communication changes dynamically during communication and is periodically regenerated, so the risk of replay attacks and man-in-the-middle attacks is greatly reduced.

[0032] Furthermore, the anomaly detection module monitors abnormal communication patterns, and if unusual access or unexpected behavior is detected, additional authentication procedures (such as multi-factor authentication) can be applied immediately. In this way, the security level of the entire communication system is dynamically managed, enabling real-time defensive responses. Example 2: Application to an unmanned aerial vehicle (drone) operating system

[0033] Unmanned aerial vehicles (UAVs), such as military drones, receive instructions via remote control and perform designated missions, but they are always vulnerable to external tampering and hacking. The dynamic security generation system of the present invention dynamically generates authentication information used between the operating terminal and the UAV in real time to protect the operation of the UAV. The generation AI model analyzes the access data and location information of the operating terminal and generates a different authentication token for each communication session, thereby preventing unauthorized access and hacking from external sources.

[0034] If the anomaly detection module detects an unauthorized access pattern during operation, the operating terminal will be required to go through an additional authentication process. In this way, the risk of unauthorized operation of military drones is reduced, and safety during mission execution is ensured. In particular, dynamic authentication using generative AI allows for a rapid response to unpredictable attacks, thereby improving the reliability of the entire system. Example 3: Application to a missile defense system

[0035] Missile defense systems are critical military systems that defend a country from enemy attacks, making security extremely important. The dynamic security generation AI system of the present invention is applied to the generation of authentication information within missile defense systems, preventing the disablement of the defense system due to unauthorized access and the risk of external manipulation. The generation AI model generates a one-time authentication token for each missile launch and interception order, based on device information such as the defense system's operating terminals and radar, and behavioral patterns.

[0036] While the defense system is operating in real time, the anomaly detection module continuously monitors access patterns and performs additional multi-factor authentication if there are signs of unauthorized access. This ensures that the missile defense system's operations and instructions are not tampered with, and that the system always operates with high reliability. Furthermore, because each token is dynamically generated by the generative AI, it is difficult for adversaries to decrypt the encryption in real time, dramatically improving the system's security.

[0037] From the above embodiments, it is confirmed that the dynamic security generation AI system of the present invention is applicable to military defense systems in general and improves responsiveness and reliability against various threats. The system of the present invention meets the high security standards required for military defense and dramatically enhances the ability to defend against cyberattacks and unauthorized access. [Industrial applicability]

[0038] The dynamic security generation AI system of the present invention is an innovative technology specifically designed for cybersecurity in military defense systems and is expected to have a wide range of applications in various military industries and related defense infrastructure. In particular, it can dramatically enhance the responsiveness and defensive capabilities against unauthorized access, information leaks, and cyberattacks from external sources in diverse military defense fields such as unmanned aerial vehicles (drones), defensive communication systems, and missile defense systems.

[0039] First, the dynamic generation and management of authentication tokens using generation AI overcomes the shortcomings of conventional security systems that rely on static authentication information, enabling flexible and highly reliable security responses to threats that change in real time. The system of this invention will significantly improve the security of communications and operations in the military industry, making it a crucial technological asset for defense-related companies that require the latest security measures.

[0040] Next, the anomaly detection module of the present invention can detect suspicious changes in access patterns and abnormal behavior in real time, providing seamless authentication during normal access while automatically raising the security level when a cyberattack is suspected. This function significantly reduces the risk of external interference to operational activities and defense system operations in military facilities and battlefield environments, making it highly practical for the defense industry in various countries.

[0041] Furthermore, by applying ensemble learning technology using generative AI, the risk of hallucination is reduced, enabling the provision of secure authentication information even to military communications and command and control systems that require high levels of security. The system of this invention is a technology that can play a strategically important role in defense-related companies and government agencies around the world, as it enables highly reliable security management while possessing the capability to defend against hostile cyberattacks.

[0042] Furthermore, since the dynamic security generation AI system of the present invention is provided on a subscription basis, it continuously receives the latest security updates and feature additions, enabling a rapid response to the latest cyber threats at all times. This model reduces the burden of security maintenance and allows for the maintenance of up-to-date security over the long term, which is expected to reduce costs and improve management efficiency for security management in defense facilities and military organizations.

[0043] The technology of this invention can be applied to industrial fields other than military in the future. For example, in fields requiring high levels of security, such as financial institutions, medical institutions, and government agencies, the use of the dynamic security generation AI system of this invention can reduce the risk of unauthorized access and information leakage, and enable real-time anomaly detection and dynamic security level management. Furthermore, as the number of IoT devices increases, it is also effective as a technology to individually protect the security of each device, making it applicable to a wide range of industries.

[0044] Based on the above, the dynamic security generation AI system of the present invention provides a new method for real-time security protection and prevention of unauthorized access in various fields, primarily the defense industry, and can be said to have very high industrial applicability.

Claims

1. A dynamic security generation system characterized by using a generative AI model to analyze user behavior data, device information, location information, and access time in real time, and generating one-time-use authentication tokens for use in military defense systems, thereby preventing unauthorized access and manipulation.

2. The dynamic security generation system according to claim 1, characterized in that the generating AI dynamically generates authentication tokens to protect the communication protocols and control signals of attack drones, block malicious control signals from hostile drones, and prevent control of the system.

3. The dynamic security generation system according to claim 2, characterized in that the generating AI model generates a different authentication token for each communication session in order to protect communication between the attack drone's operating terminal and the drone itself, thereby preventing unauthorized operation and hacking from external sources.

4. A dynamic security generation system according to claim 2 or 3, characterized in that an anomaly detection module monitors the operation patterns of an attacking drone, and automatically blocks communication when unauthorized access or abnormal operation is detected, thereby preventing unauthorized operation of the drone.

5. The dynamic security generation system according to claim 1, characterized in that authentication tokens are generated in real time to prevent interference between the defensive drone and adversarial drones or unmanned aerial vehicles (UAVs), and communication between drones is securely protected.

6. The dynamic security generation system according to claim 5, characterized in that the generating AI model analyzes the operation of the defense drone and applies additional multi-factor authentication (MFA) to raise the security level when abnormal access or suspicious control signals are detected.

7. A dynamic security generation system according to claim 5 or 6, characterized in that communication between the defense drone and the command terminal is protected by encryption technology using TLS (Transport Layer Security) or SSL (Secure Sockets Layer), and external access is immediately blocked.

8. A dynamic security generation system according to claim 1, characterized in that it generates authentication tokens used for instructing and controlling a missile defense system in real time, thereby preventing unauthorized operation of the system or unauthorized instructions from external sources.

9. A dynamic security generation system according to claim 8, characterized in that the generated AI model monitors for unauthorized access targeting the operating terminals and radar systems of a defense system, and requests additional authentication when an anomaly is detected.

10. A dynamic security generation system according to claim 8 or 9, characterized in that it identifies multiple access terminals in a missile defense system and dynamically generates authentication information in real time to prevent unauthorized use of each access terminal. Claims relating to common functions

11. The dynamic security generation system according to claim 1, characterized in that the generating AI model reduces the risk of misgenerating authentication tokens based on ensemble learning technology, and if misgeneration occurs, other models correct it, thereby providing accurate and reliable authentication information.

12. A dynamic security generation system according to any one of claims 1 to 11, characterized in that an anomaly detection module detects unauthorized access or suspicious behavior in real time, and if detected, immediately raises the security level and performs additional authentication or communication blocking.

13. A dynamic security generation system according to any one of claims 1 to 12, characterized in that it cooperates with an external authentication server via an API using encryption technology such as TLS (Transport Layer Security) or SSL (Secure Sockets Layer), and authentication information is communicated securely.

14. A dynamic security generation system according to any one of claims 1 to 13, characterized in that an authentication token is securely transmitted over a communication path after generation and is automatically invalidated after use, thereby preventing unauthorized reuse.

15. A dynamic security generation system according to any one of claims 1 to 14, characterized in that the generation AI is provided in a subscription format, periodic security updates are applied in real time, and the system always maintains the latest security status.

16. A dynamic security generation system according to any one of claims 1 to 15, characterized in that when a cyberattack is detected, countermeasures are generated in real time, and additional authentication or communication blocking is performed immediately to ensure the security of the system.

Images