The essentials
Multiple AI safety researchers resign from leading companies, warning of existential risks from uncontrolled AI advancement.
Jacob Coxon, an Anthropic researcher, has resigned citing concerns that AI could pose an existential threat by the end of the decade. Coxon previously worked on OpenAI's GPT-4o model from 2023 to July 2026 before joining Anthropic, providing him experience across two major AI development organizations.
Coxon describes different risk approaches between OpenAI and Anthropic, highlighting divergent safety philosophies in the industry. Evan Hubinger, an Anthropic safety researcher, has supported Coxon's concerns about AI extinction risk, signaling internal acknowledgment of safety challenges within the organization.
Coxon's resignation is part of a broader pattern of researcher departures focused on safety concerns. Mrinank Sharma, an Anthropic safety researcher, resigned in February 2026, and OpenAI researcher Hieu Pham resigned in February 2026 citing AI existential threat. OpenAI alignment chief Jan Leike departed in 2024 citing safety concerns, establishing an earlier precedent for safety-focused departures.
These individual resignations align with broader industry concerns. Over 1,300 AI company employees signed an open letter warning of uncontrolled capability acceleration, reflecting widening anxiety about AI development trajectories across the sector.
Other ways to read this story.
Alternative framings generated by SigNews. These are not independent sources or attributed quotations.
Safety researchers departing AI companies amid existential risk warnings
Four documented researcher departures from OpenAI and Anthropic between 2024-2026 reveal a pattern of safety-focused exits. Jacob Coxon, Mrinank Sharma, Hieu Pham, and Jan Leike all departed citing safety and existential risk concerns. These departures span different organizations and timeframes, suggesting systemic concerns rather than isolated disagreements about company direction.
Broader workforce alignment with safety concerns across AI sector
Over 1,300 AI company employees signed an open letter warning of uncontrolled capability acceleration, extending concerns beyond individual researchers at specific organizations. This broad-based support demonstrates alignment across the workforce on capability risks. Individual researcher departures coincide with industry-wide expressions of concern about development velocity and inadequate safety mechanisms.