Anthropic Deployed Six Engineers to NSA for Offensive Cyber Operations Using Mythos AI, Financial Times Reports Thursday

According to Financial Times, Anthropic deployed approximately six engineers to the National Security Agency to help deploy its Mythos AI model for offensive cyber operations, potentially including attacks on networks in China and Iran. The engineers are customizing the model for specific applications as forward-deployed staff.

On the same day, Anthropic's research institute published a report warning that Claude has automated over 80% of code merged into its production codebase—up from low single digits before early 2025—and called for a verifiable global pause in AI development. The report cited an April experiment where Claude agents recovered 97% of a performance gap in an AI safety problem, compared to 23% recovered by human researchers over a week.