Bitmax Launches Autonomous AI Security Platform 'PatchGuard': "AI Attacks Require AI Defense"

Bitmax, a company engaged in system integration (SI) business, announced on May 11 that it is expanding its business into the artificial intelligence (AI)-based security market with the official launch of its autonomous security platform, PatchGuard.

The company explained that this launch demonstrates how AI-driven cyber security threats have moved beyond being a future possibility and have become a present and realistic danger.

Recently, in the global AI industry, Anthropic has drawn attention for not making its latest model, Claude Opus 4 (Mythos), generally available, citing "the possibility that its powerful performance could be exploited for security threats." Concerns have been raised that the model's exceptional ability to detect software vulnerabilities could be used in large-scale cyberattacks.

A Bitmax representative stated, "With AI-powered cyberattacks already a reality, we have confirmed that even without public release of specific models, advanced attack scenarios can be implemented solely through prompt engineering and fine-tuning."

However, Bitmax clarified that it has applied an agentic AI architecture based on existing open AI models, not Mythos.

Along with the launch of PatchGuard, Bitmax also conducted security assessments of 30 major domestic websites using its proprietary deep-diagnostic engine, URL Scan Flash 1.0. The assessment found that 28 sites, or 93% of the total, contained critical vulnerabilities that could be exploited for hacking, and not a single site fully met the ISMS-P (Personal Information & Privacy Management System) standards.

In addition, the detection rate for personal information protection-related vulnerabilities reached 97%, indicating a high risk of exposure to phishing and data theft attacks. Critical vulnerabilities were found across key industries, including healthcare, universities, financial sector sites, as well as e-commerce and public institutions, and overall, they were rated as being at a "very dangerous" level.

Notably, the majority of the vulnerabilities identified were found to stem from basic security misconfigurations rather than more complex hacking techniques such as session hijacking, authentication bypass, man-in-the-middle (MITM) attacks, subdomain takeover, or API exposure. The company explained that as AI is now capable of automatically identifying such weaknesses, the limitations of traditional, manual inspection-based security systems are becoming increasingly apparent.

The core technology of PatchGuard is its "Deep Hunt" security detection engine, which employs an agentic AI structure. This engine enables AI to autonomously scan attack surfaces, identify vulnerable points, and repeatedly verify them. Unlike traditional rule-based and manual security solutions, this multi-layered detection system features automatic identification of attack surfaces, priority-based deep scanning, autonomous verification in a sandbox environment, and vulnerability determination through repeated verification.

In addition, PatchGuard is equipped with an "AI Repair" function that automatically generates, tests, and verifies correction code, thereby automating the entire process from detection to response and recovery. This enables rapid response to high-risk vulnerabilities that were difficult to detect using conventional methods.

A Bitmax representative commented, "As attackers now use AI to find vulnerabilities at unprecedented speed, the era of manually inspecting security is over. It is now essential to adopt an automated system where AI detects and defends against vulnerabilities on its own."

An industry expert also emphasized, "Responding to AI security threats should not focus solely on whether specific models are made public, but rather on detecting and blocking the misuse of widely distributed AI models."

Meanwhile, Bitmax currently offers a free scanning service that allows users to check basic security status simply by entering a URL. However, considering the potential for misuse, AI-based deep analysis features such as URL Scan Flash and repository integration will be released gradually.

As AI-driven cyberattacks become a reality, Bitmax is working to build a next-generation security system that automates the entire process from detection to response and recovery through its autonomous AI security platform, PatchGuard.

This content was produced with the assistance of AI translation services.