AL
Alignment Forum
1 story this week · 6 topics · alignmentforum.org
Research forum for technical AI safety and alignment work.
Latest from Alignment Forum
- Public SectorSetting the Scene (Public v1; October 25)
- Society & CultureThe Alignment Problem: Computer Science Meets Analytic Philosophy
- AI ResearchA Simpler and Better Debate Protocol for Scalable Oversight
- Society & CulturePreface to the Computational Agent Theory and ASI Alignment Sequence
- AI ResearchThe Continual Constitutional Convention: How language models revise the documents that govern them
- Society & CultureOpenAI Can't Self-Police Because It Fires the People Raising the Alarm
- Machine LearningMinimum required pacing: finish aligning it before you train the next one
- Society & CultureA Defense of Gradual Disempowerment
- AI ResearchThe Alignment Journal: Organization, Personnel, and Scope
- AI ResearchTraining a Misaligned Reward Seeker
- AI ResearchValue generalisation Theory of Change: putting it into practice
- AI ResearchValue generalisation Theory of Change: the theory behind the approach
- AI ResearchBrief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident
- AI ResearchDebate Training Reduces Reward Hacking in RLAIF
- AI ResearchAI swarms are starting to pose indirect takeover risk
- AI ResearchFour LLM loss functions → four flavors of LLM misalignment
- AI ResearchWhy do models task game?
- AI ResearchReturning to ARC
- AI ResearchAGI Safety and Alignment at Google DeepMind: A Summary of Recent Work (July 2026)
- AI ResearchThe AGI Safety and Alignment team at Google DeepMind is Hiring (July 2026)
- AI ResearchThousand-dimensional structure
- AI ResearchValue Generalisation 3: Pre-aligned AIs
- AI ResearchValue Generalisation 2: The Missing Hole in AIs’ abilities
- AI ResearchValue Generalisation 1: a Research and Deployment Program
- AI ResearchResearch directions in condensation: varieties of objectivity
- AI ResearchRL & search is a terrifying way to build AGI (an FAQ)