Ethical AI Departures

A sourced tracker of researchers, engineers and executives who left AI companies over safety or ethics concerns.

Links

• Site: https://ethicalaidepartures.fyi/
• About and inclusion criteria: https://ethicalaidepartures.fyi/about

Overview

The tracker lists 46 departures (as of October 2026) from OpenAI, Google, Anthropic and others. Each record is graded by how strong the evidence is. Direct means the person linked the exit to the concern themselves. Reported means independent reporting established the link. Alleged cases are disputed or unresolved, and are listed separately and left out of the count. It also records the disbanding of safety teams separately, as organisational events. The latest entry is David Robinson's September 2026 resignation from OpenAI over its "launch-driven, trial-and-error" approach to safety.

It is run by one person, is independently funded and offers email alerts for new profiles. The site frames safety-motivated departures as "a leading indicator of organizational culture around AI risk."

Related

• moc-ai-risk-warnings
• bengio-leave-frontier-ai-companies: Bengio urges safety-minded researchers to leave the labs