Machine Intelligence Research Institute (MIRI)

The Machine Intelligence Research Institute (MIRI) is a nonprofit research organization dedicated to ensuring that the development of artificial general intelligence (AGI) benefits humanity. Founded in 2000 as the Singularity Institute for Artificial Intelligence, MIRI conducts research on AI alignment, decision theory, and formal verification—seeking mathematical and logical approaches to the problem of ensuring advanced AI systems behave as intended.

MIRI's research agenda emphasizes theoretical foundations rather than applied AI development. Key research areas include logical uncertainty, decision theory (particularly Functional Decision Theory), agent foundations, and corrigibility. The organization has published numerous papers in academic venues and maintains a research blog discussing AI safety topics.

Based in Berkeley, California, MIRI has become a central institution in the AI safety field, influencing academic discourse and attracting researchers concerned with long-term AI risks. The organization is supported by donations from individuals and foundations interested in existential risk mitigation.