Cyber security experts are beginning to warn that AI-powered attacks are increasing their threat as Anthropic developed its Mythos model, which can detect security vulnerabilities at a faster rate than human hackers.
This threat impacts governments, corporations, and critical infrastructure around the world, raising concerns about protecting against AI-powered exploits in real-time.
According to a leaked blog post, AI developer Anthropic has warned that its upcoming Mythos model will conduct automated software vulnerability scanning and exploitation at extraordinary speed. OpenAI issued a similar warning in December about future models posing higher cybersecurity risks.
Shlomo Kramer, founder and CEO of Cato Networks, commented: “Aggressive attackers are coming. This is a significant event in the history of cybersecurity.” According to experts, an AI agent can defeat hundreds of human hackers by its constant scanning, testing and exploiting vulnerabilities.
A Russian-speaking hacker used Anthropic’s Cloud and a variety of AI tools, such as China’s DeepSeek, to hack more than 600 devices in 55 countries in January. According to Amazon Web Services’ security research team, this was done using AI tools. According to Eyal Sela, director of threat intelligence at Gambit Security, AI has made complex hacks easier for attackers, giving them “superpowers” regardless of their skills.
Cloud was also reportedly used in attacks on Mexican government agencies in February, where sensitive voter and tax data were leaked.
AI can also be used to improve cyber security. Joe Lin, co-founder and CEO of Twenty, commented: “Attackers only need one path in, while defenders must cover every surface. AI enables faster monitoring and automated patching, but humans must control the outcomes.”
Experts highlight how AI-based cybersecurity tools will be essential in combating the growth of autonomous AI attacks.
Anthropic and other AI development centers are testing their models with select organizations to prepare for large-scale AI-based cyber threats.
