Machine Intelligence Research Institute

The Machine Intelligence Research Institute (MIRI) is a nonprofit research organization based in Berkeley, California that focuses on long-term AI safety, particularly the challenge of ensuring that highly capable future AI systems behave in ways aligned with human values. Founded in 2000 (originally as the Singularity Institute for Artificial Intelligence), it has been influential in shaping the early field of AI alignment research and remains one of the best-known organizations focused on existential risks from advanced AI. Wikipedia

Mission and research

MIRI's stated mission is to reduce the risk that advanced artificial intelligence could cause catastrophic outcomes. Its current public position is that creating artificial superintelligence using approaches similar to today's techniques poses an extremely high risk, and that delaying such development is preferable until safer methods are understood.

Its research has emphasized foundational theoretical work rather than building AI systems directly. Topics have included:

• AI alignment and value alignment
• Decision theory
• Agent foundations
• Logical uncertainty
• Mathematical frameworks for reasoning about advanced AI systems

These topics aim to establish rigorous theoretical foundations for designing AI systems whose objectives remain aligned with human intentions.

Historical significance

MIRI is widely recognized as an early contributor to AI safety as a research area. Its researchers and collaborators helped popularize concepts such as AI alignment and long-term existential risk from artificial general intelligence, years before these topics became mainstream within academia and industry. The organization has also produced educational materials and public outreach intended to encourage more research on these problems.

Perspectives and debate

MIRI's outlook is among the more cautious—and in some respects, the most pessimistic—within the AI safety community. It argues that sufficiently powerful AI developed with current paradigms is likely to pose existential risks if deployed before major theoretical breakthroughs in alignment are achieved. Machine Intelligence Research Institute

This position is influential but not universally shared. While many AI researchers agree that advanced AI safety deserves serious attention, there is substantial disagreement about:

• how likely catastrophic outcomes are,
• how soon highly capable AI systems may emerge,
• and which technical or governance strategies are most effective for reducing risk.

Organization

MIRI is a U.S. 501(c)(3) nonprofit headquartered in Berkeley, California. It conducts research, publishes technical work, and engages in public communication about advanced AI safety and alignment. The organization was renamed from the Singularity Institute for Artificial Intelligence in 2013.