AI Safety Crisis: Anthropic Lead Resigns

Anthropic Safeguards Lead Resigns, Citing AI Safety Concerns

• Mrinank Sharma, former head of safeguards research at Anthropic, resigned and shared his departure letter publicly.
• Sharma expressed concerns over the disconnect between ethical commitments and operational practices in AI organizations.
• He worked for two years on defenses against AI-enabled biological threats and internal accountability tools.
• Sharma plans to pursue writing, personal coaching, and possibly graduate studies in poetry after leaving corporate AI work.
• His resignation highlights growing scrutiny on how AI developers handle internal dissent and balance capability gains with safety research.

Mrinank Sharma has left Anthropic due to concerns about the gap between ethical principles and practical decisions in AI organizations. His departure underscores increased attention on managing risks and internal dissent within the industry.