22 evidence-linked departures · 5 unresolved · 1 organizational event
OpenAI has the largest set of evidence-linked records in this tracker. They include Jan Leike, who said safety had taken 'a backseat to shiny products' after the Superalignment team dissolved; Daniel Kokotajlo, who said he lost confidence that the company would act responsibly around AGI; and Miles Brundage, who warned that 'neither OpenAI nor any other frontier lab is ready' for AGI. The record continues into 2026 with departures tied by public statements or independent reporting to ChatGPT advertising plans and a Pentagon partnership. The Superalignment team's dissolution is documented separately as an organizational event and does not establish any other individual's motive.
David Robinson
Safety Transparency Lead, Safety Systems · 2026
Caitlin Kalinowski
VP of Hardware, Robotics Lead · 2026
Hieu Pham
Member of Technical Staff · 2026
Zoe Hitzig
Researcher · 2026
Pamela Mishkin
Economics research team lead · 2026
Team-level changes are documented separately and never establish an individual's motive or enter the departure count.
OpenAI dissolved the Superalignment team less than a year after announcing it as a dedicated effort to control AI systems smarter than humans.
OpenAI said the work would be integrated into other research efforts. The dissolution followed the departures of co-leads Ilya Sutskever and Jan Leike.
People and motive: Team membership or departure timing does not establish why any other individual left. Only people with person-level evidence appear in the departure tally.
These claims are disputed or unresolved and are excluded from the primary tally.
Leopold Aschenbrenner
Researcher, Superalignment Team · 2024