Machine Intelligence Research Institute

What MIRI is

The Machine Intelligence Research Institute (MIRI) is a 501(c)(3) nonprofit based in Berkeley, California, which carries out research and public outreach intended to help prevent human extinction from the development of artificial superintelligence^[2]. It describes itself as a non-profit research institute focused on identifying and managing potential existential risks from artificial intelligence^[1].

Origins

MIRI began in 2000, when Eliezer Yudkowsky founded the Singularity Institute for Artificial Intelligence with funding from Brian and Sabine Atkins. The original purpose was not risk reduction but the opposite: accelerating the development of artificial intelligence^[1]. MIRI's own account of its history also places the founding in 2000, names Yudkowsky together with Brian and Sabine Atkins as founders, and gives the same original goal of accelerating progress towards smarter-than-human AI^[2].

In 2005 the institute began operating in Silicon Valley and turned its attention to ways of identifying and managing AI risks, a topic that at the time was largely ignored by scientists in the field^[1]. Starting in 2006 it organized the Singularity Summit, a conference on the future of AI, initially in cooperation with Stanford University and with funding from Peter Thiel^[1].

The name change

In December 2012 the institute sold its name, its web domain and the Singularity Summit to the Singularity Group, and took the name Machine Intelligence Research Institute the following month^[1]. MIRI's official history dates the change of name from the Singularity Institute to the Machine Intelligence Research Institute to 2013^[2]; the two accounts differ on when the transition is dated.

Research

MIRI's earlier research, led by Yudkowsky, primarily addressed the AI control problem and the question of how to design friendly AI^[1]. The institute has also funded forecasting work through an initiative called AI Impacts, which studies historical instances of discontinuous technological change^[1]. MIRI describes itself as the first organization to advocate for and work on alignment of artificial superintelligence as a technical problem^[2].

Funding

In 2019 Open Philanthropy recommended a general-support grant of approximately $2.1 million over two years to MIRI^[1]. In April 2020 MIRI announced an additional $7.7 million grant from Open Philanthropy, also over two years^[1]. In 2021 Vitalik Buterin donated several million dollars' worth of Ethereum to the institute^[1].

A shift in priorities

By 2023–2024 MIRI had concluded that the default consequence of further AI development would be human extinction, and it began to prioritize technical AI governance and public communication^[1]. In January 2024 it announced that it had shifted its priorities in 2023 toward policy and public communication, advocating an international agreement to halt progress toward smarter-than-human AI^[1]. Its stated priority is to inform the world about the situation and to advocate stopping AI development altogether, until such a time as the alignment problem has been solved^[2]. In 2025 co-founder Eliezer Yudkowsky and president Nate Soares published the book "If Anyone Builds It, Everyone Dies"^[1].

Where it sits in the field

MIRI's trajectory runs from a project meant to speed AI up, through a long stretch of technical work on control and alignment, to an organization that argues that development should be paused. Its early framing of AI risk as an engineering problem, at a time when the surrounding field paid little attention to it^[1], is the reason it is usually treated as a starting point for later work on AI safety^[2].