Claude Mythos: AI Hacking Tool and Its Risks
Tuesday, 2026/04/21144 words2 minutes209 reads
Anthropic, a leading AI company, has developed Claude Mythos, a new AI model with exceptional hacking capabilities. The tool has demonstrated the ability to outperform humans at certain cyber-security tasks, particularly in finding hidden vulnerabilities in computer code.
During testing, researchers found that Mythos could locate dormant bugs in decades-old systems and easily exploit them. In one case, it discovered a vulnerability that had existed for 27 years. Due to these powerful capabilities, Anthropic decided not to release Mythos to the general public.
Instead, the company launched Project Glasswing, giving access to 12 major tech companies including Amazon, Apple, Microsoft, and Google. The goal is to help these organizations strengthen their defenses against potential threats. However, finance ministers and central bankers have expressed serious concerns about the risks Mythos could pose to financial systems and critical infrastructure if it falls into the wrong hands.
