The Center for AI Safety recognizes that building beneficial artificial intelligence is more than a theoretical problem. It requires mass coordination and systemic intervention. To this end, we work with stakeholders across the spectrum to reduce risk from large-scale AI systems. We support projects at all stages in the pipeline, from basic research to deployment/monitoring tools to cultural interventions.
Some of our work involves setting technical research directions and distributing the resources required to make progress in these directions. We fund educational efforts that redirect focus from AI capabilities to AI social impact and alignment with human values. We manage relationships between technical parties and policymakers. Our goal is to empower those who will shape the future of AI so that they may do so safely.
