Anthropic Safeguards Head Resigns Over Values Clash

Mrinank Sharma, the head of the Safeguards Research Team at Anthropic, resigned from the San Francisco-based artificial intelligence company on February 9, 2026, citing internal pressures to prioritize commercial interests over core values and broader global crises threatening humanity.

In a letter shared with colleagues and posted on the social media platform X, Sharma described his two years at Anthropic as meaningful, highlighting achievements such as studying AI sycophancy, the tendency of systems to overly agree with users; developing defenses against AI-assisted bioterrorism, some of which were deployed in production; and authoring one of the first AI safety cases. He also worked on internal transparency mechanisms and explored how AI assistants might distort human judgment or reduce empathy. "I've achieved what I wanted to here," Sharma wrote, expressing gratitude for the team's intellectual rigor and kindness.

However, Sharma's departure stemmed from mounting concerns about the misalignment between ethical commitments and operational realities. "Throughout my time here, I've repeatedly seen how hard it is to truly let our values govern our actions," he stated in the...