Anthropic found Mythos AI too dangerous

Anthropic's internal AI experts identified significant risks associated with their AI model, Mythos, warning that it could potentially compromise critical systems underpinning modern computing infrastructure. This internal assessment highlights the inherent dangers and potential vulnerabilities that advanced AI technologies can introduce. Financial institutions and government agencies are now reportedly accelerating their evaluations of such threats, indicating a heightened awareness and proactive stance towards mitigating the risks posed by powerful AI systems.