Model welfare
Research on whether AI systems could be moral patients, and what follows.
36 documents, newest first
Wellbeing research grantsProbing persona-dependent preferences in language modelsInitial results of the Digital Consciousness ModelWhere is the mind? Persona vectors and LLM individuationInternal experience machinesArtificial Intelligence, Morality, and Sentience (AIMS) SurveyThe evidence for AI consciousness, todayShow Notes: Eleos on Bloomberg's Odd LotsAI and ConsciousnessWhy it makes sense to let Claude exit conversationsSome thoughts on human-AI relationshipsWhy model self-reports are insufficient—and why we studied them anywayEleos commends Anthropic model welfare effortsResearch priorities for AI welfareAI welfare organization Eleos expands team with hires from OpenAI and OxfordPrinciples for responsible machine consciousness research (open letter)Key concepts and current beliefs about AI moral patienthoodWorking paper: key strategic considerations for taking action on AI welfarePrinciples for Responsible AI Consciousness ResearchWorking paper: review of AI welfare interventionsStrategic Directions for a Digital Consciousness ModelInsects, AI systems, and the future of legal personhoodThe Welfare of Digital MindsTaking AI Welfare SeriouslyNew report: Taking AI Welfare SeriouslyNew AI welfare report coming soonExperts Who Say That AI Welfare is a Serious Near-term PossibilityRethink Priorities’ Digital Consciousness Project AnnouncementPrioritizing animals of uncertain sentienceCould a Large Language Model Be Conscious?Artificial Intelligence, Morality, and Sentience (AIMS) Survey 2023Do Brains Contain Many Conscious Subsystems? If So, Should We Act Differently?Theories of Welfare and Welfare Range EstimatesAn Introduction to the Moral Weight ProjectPropositions Concerning Digital Minds and SocietySharing the World with Digital Minds