Context Week
Context Week was a four-day residential experiment about AI safety. The experiment has concluded.
What we did
We brought together 24 participants at Lighthaven in Berkeley for four programme days, from 31 August to 3 September 2026. The group included people exploring AI safety and people already working in the field.
Participants read primary sources on AI capabilities and risks, discussed them in small groups, and took part in debates with assigned positions. They wrote about desirable futures, compared strategies for responding to advanced AI, and discussed organisations working on AI safety.
The week also included career writing, feedback on participants' plans, and one-to-one conversations with facilitators, practitioners and peers. Participants drafted career memos setting out their options, questions and possible next steps.
Express interest
You can still fill out the application form as an expression of interest, but you may not hear back.