AI Safety Lead Quits Anthropic After Two-Year Tenure Citing Global Perils
Anthropic’s AI safety lead Mrinank Sharma resigned on Monday after spearheading defenses against AI-driven bioterrorism and developing internal transparency mechanisms since August 2023. Amazon, a key investor in the Claude developer, faces heightened scrutiny as Sharma warns of “interconnected crises” posed by advanced AI models.
1. Resignation Details
Mrinank Sharma, who joined Anthropic in August 2023 as head of AI safety, submitted his final day on Monday via a letter warning of “interconnected crises” posed by AI and bioweapons. He stated he will pursue work more aligned with personal values and integrity.
2. Sharma's Safety Work
Sharma spearheaded research on AI sycophancy, defenses against AI-assisted bioterrorism, and transparency mechanisms within Anthropic’s systems. He described these initiatives as his core contributions since joining.
3. Amazon's Stake and Risks
Anthropic has raised billions from Amazon and Alphabet, making Amazon a significant backer of the Claude chatbot developer. Sharma’s exit may heighten investor and regulatory scrutiny as competition and safety concerns intensify in the AI sector.