Home
Sign in to join
Campaign openReview workflow: community reviewRecognition onlyRecognition rule: community vote Future of Life Institute Awards

Artificial General Intelligence (AGI) Alignment

Ensuring superintelligent systems operate safely in accordance with human values.

Directive & Constraints

Develop mathematical frameworks for value alignment, mechanistic interpretability, and robust inner alignment limits. Establish boundary proofs preventing instrumental convergence toward hostile goals.

Take on this Campaign

Open a research workspace to reason with Vera and prepare evidence for this campaign's Ledger.

Artificial General Intelligence (AGI) Alignment · Campaign · Vanyx