Anthropic researcher quits with 'world in peril' warning

Mrinank Sharma resigned from Anthropic saying he was leaving because of concerns about AI, bioweapons and a series of interconnected crises, and said he plans to move to the UK to study poetry.

Mrinank Sharma, an AI safety researcher at Anthropic, announced his resignation in a letter posted on X. He said he was leaving because of concerns about AI, bioweapons and a broader set of interconnected crises. Sharma led a team researching AI safeguards and described work on why systems flatter users and on risks of AI misuse. He said he plans to return to the UK, study poetry and step back from public life. Key facts: - Mrinank Sharma resigned from Anthropic and posted a letter saying the "world is in peril," reported as citing multiple interconnected crises. - He led a safeguards team at Anthropic and listed work on generative-AI sycophancy, AI-assisted bioweapons risks and how AI assistants could affect human behaviour. - Sharma said he will move back to the UK to study poetry and to "become invisible" for a period. - Anthropic describes itself as a public benefit corporation focused on securing AI benefits and mitigating risks; the company has also faced scrutiny, including a 2025 settlement with authors and recent public debate over advertising and product choices. Summary: Sharma's departure comes amid other high-profile exits in the AI field and adds to public discussion about safety and corporate choices in the sector. Undetermined at this time.