Anthropic has revealed that approximately 67% of the 832 accounts banned over the past year for violations utilized AI to aid in cyberattacks, including malware creation. The data highlights the increasing role of AI in the attack lifecycle, with 6.5% of these accounts employing AI for internal network penetration. The proportion of medium- to high-risk accounts surged from 33% to 56% over the year. In a related development, Google identified the first instance of AI being used to develop a zero-day vulnerability last month. Anthropic is preparing to launch its AI model, Mythos, which has already identified over 10,000 critical vulnerabilities in widely used software, underscoring the dual-use nature of AI in cybersecurity.