Mrinank Sharma, who led Anthropic's Safeguards Research team, resigned with a stark warning about the state of the world, stating it is "in peril" not just from AI and bioweapons, but from a series of interconnected crises. Sharma, who holds a Ph.D. in machine learning from the University of Oxford and joined Anthropic in August 2023, shared his resignation letter on X, where it quickly gained significant attention, amassing over 1 million views.
In his letter, Sharma indicated that despite achieving his goals at the company, including developing defenses against AI-assisted bioterrorism and studying AI's effects on humanity, it was time to move on. He expressed concern that it is difficult "to truly let our values govern our actions" and implied that Anthropic, like other companies, faces "pressures to set aside what matters most." Anthropic, a "public benefit corporation," aims to secure AI's benefits and mitigate risks, focusing on preventing advanced AI systems from becoming misaligned with human values or misused.
Following his departure, Sharma announced plans to move back to the UK, pursue a poetry degree, and "become invisible" for a period. His exit comes amidst a trend of high-profile AI researchers and executives leaving major AI firms like Anthropic and OpenAI, often sounding alarms about the technology's potential risks on their way out. Other recent departures from Anthropic include Harsh Mehta, Behnam Neyshabur, and Dylan Scandinaro, who moved to OpenAI to head its preparedness team.
This wave of departures occurs as leading AI companies are reportedly racing towards potential IPOs and seeking new funding rounds, which could significantly boost their growth while also attracting increased scrutiny. For instance, Anthropic was in talks to raise a new funding round and had previously settled a $1.5 billion class-action lawsuit in 2025 with authors who claimed their work was used to train AI models without consent. Sharma's resignation underscores growing concerns within the AI community about ethical considerations, safety, and the commercial pressures influencing AI development.
