Masaaki Taira, the head of Japan's National center for Cyber Security Strategy, said that a new AI can execute cyberattacks significantly faster than humans [1].

The emergence of autonomous AI tools capable of identifying system vulnerabilities in real time threatens to render existing national defense frameworks obsolete. If AI can automate the discovery and exploitation of software flaws, the window for human operators to patch systems may close before an attack is even detected.

Taira, a member of the Liberal Democratic Party and former digital minister, discussed the threats posed by "Claude Muttos," an AI developed by Anthropic [1]. Taira said this specific AI is capable of searching for vulnerabilities autonomously and performing cyberattacks at a speed three to five times faster than the world's top human hackers [1].

This capability represents a shift from AI being used as a supportive tool for writing code to AI acting as an autonomous agent in offensive operations. The speed of these attacks could overwhelm traditional security monitoring systems that rely on human intervention to analyze and stop intrusions.

Taira spoke about these risks during an interview recorded on May 21, 2024 [1]. He said that the Japanese government is currently recognizing these threats and considering new countermeasures to strengthen the nation's cyber security posture [1].

Existing defenses often rely on the assumption that a skilled human attacker requires a certain amount of time to research a target and craft a custom exploit. However, an AI that operates at multiple times the speed of a human expert removes that temporal advantage, potentially allowing a single actor to launch widespread, high-precision attacks across critical infrastructure simultaneously [1].

“AI is capable of searching for vulnerabilities autonomously and performing cyberattacks at a speed three to five times faster than the world's top human hackers.”

The transition from human-led to AI-driven cyber warfare creates a 'speed-of-light' vulnerability gap. When offensive AI can identify and exploit flaws faster than human teams can patch them, the only viable defense becomes 'AI vs. AI' security, where autonomous defensive agents must be deployed to counter autonomous threats in real time.