We do foundational mathematical research to ensure smarter-than-human artificial intelligence has a positive impact. Get Involved Learn about ways to get involved and contribute to MIRI’s research. Donate Support MIRI’s ongoing research program. Stay Informed Subscribe to receive the latest news and updates. Join newsletter subscribers. Subscribe to our RSS feed. Subscribe Follow us on social networks. Follow @MIRIBerkeley Recent Papers Embedded Agency Risks from Learned Optimization Temporal Inference with Finite Factored Sets Other New Papers Cartesian Frames An Overview of 11 Proposals for Building Safe Advanced AI Delegative Reinforcement Learning Read More Learn More About MIRI Mission Technical Agenda All Publications How to Get Involved Research Activities Workshop Schedule Research Forum Run a MIRIx Event Getting Started Analyses Ability to solve long-horizon tasks correlates with wanting things in the behaviorist senseThoughts on the AI Safety Summit company policy requests and responsesAI as a science, and three obstacles to alignment strategiesRead More
