OpenAI launches inaugural Safety Fellows cohort on alignment and interpretability

tomekkorbak · x · 2026-09-17

OpenAI has kicked off its inaugural Safety Fellows cohort in partnership with Constellation. First-cohort member @xaverg announced the group will spend the next few months working on core issues in AI alignment, control, and interpretability, with support from mentors and the broader community.

DeepMind researcher Tomáš Korbak shared the news, saying he expects "an unreasonable amount of ground-breaking AI safety research" to come out of this group.

Original post →

More from Companies & People

Companies & People channel →