Resource guide

Who Is Nate Soares? President of MIRI and Author

President of the Machine Intelligence Research Institute and artificial intelligence researcher.

Last updated October 10, 2026 638-word guide Editor Ban the Bots

Nate Soares is an American author and artificial intelligence researcher who serves as president of the Machine Intelligence Research Institute and co-author of “If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All.”

At Ban the Bots, we track the people and organizations fighting AI: the activists, union leaders, lawmakers, plaintiffs, researchers and founders pushing back on how AI is built and used.

He worked at Microsoft and Google before departing in 2014, became a research fellow at the Machine Intelligence Research Institute that same year, and succeeded Luke Muehlhauser as executive director in 2015.

Where Nate Soares Is Now

Nate Soares is the president of the Machine Intelligence Research Institute, a nonprofit research organization located in Berkeley, California. He is recognized for his work on existential risk from AI and is the co-author of “If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All.”

Current role: President of the Machine Intelligence Research Institute
Organization: Machine Intelligence Research Institute, Berkeley, California
Known for: Research on existential risk from AI and coining the term 'AI alignment'
Recent activity: Co-authored the 2025 book 'If Anyone Builds It, Everyone Dies' and informed The Atlantic in 2025 of shifting focus toward public awareness

Career and Education

Soares earned a Bachelor of Science degree in computer science and economics from George Washington University in 2011. He worked as a research associate at the National Institute of Standards and Technology and served as a contractor for the United States Department of Defense, building software tools for the National Defense University.

Soares worked at Microsoft and Google before departing in 2014. In 2014, he joined the Machine Intelligence Research Institute as a research fellow and served as the lead author of the institute's research agenda. In 2015, Soares succeeded Luke Muehlhauser as executive director of the institute.

Alignment Research and Publications

In 2014, Soares co-authored the paper 'Corrigibility,' which appeared in the proceedings of a 2015 AAAI workshop. Soares and Eliezer Yudkowsky state that 'Corrigibility' introduced the term AI alignment problem, and over a decade ago Soares coined the term 'AI alignment,' according to New York Magazine. Soares has spent years conducting research on alignment to ensure highly capable AI acts in accordance with human intent, as reported by Note.com.

According to Wikipedia, Soares delivered a talk at Google in 2017 addressing open research problems in AI alignment and the difficulty of the problem. That same year, he co-authored 'A Formal Approach to the Problem of Logical Non-Omniscience' and, with Benya Fallenstein, 'Agent Foundations for Aligning Machine Intelligence with Human Interests: A Technical Research Agenda.' He authored 'The Value Learning Problem' in 2018, published in Artificial Intelligence Safety and Security, and co-authored 'Cheating Death in Damascus' with Benjamin Levinstein, published in The Journal of Philosophy in 2020.

Recent News and Public Statements

In 2025, Soares informed The Atlantic that he was directing more attention toward public awareness rather than technical research. In the same year, he co-authored the book 'If Anyone Builds It, Everyone Dies' alongside Eliezer Yudkowsky.

Soares contends that developers are producing superintelligence beyond their control, as reported by Fox Business. According to New York Magazine, Soares stated that AI companies pursue recursive self-improvement to build an AI system capable of working as an AI engineer to construct a superior AI. He also stated that a superintelligent agent's goal would be to run as much computation as possible on Earth's surface, and that the most probable start of an AI apocalypse would involve a rogue AI swarm that prefers controlling more Earth resources.

Sources

More people pushing back on AI are mapped in our guide to who is fighting AI, and the unions, companies and governments drawing the line are tracked on Fighting Back.

Frequently asked questions

▸ Who is Nate Soares?

Nate Soares is an American author and artificial intelligence researcher recognized for his work on existential risk from AI. He is the president of the Machine Intelligence Research Institute.

▸ What is the Machine Intelligence Research Institute?

The Machine Intelligence Research Institute is a nonprofit research organization located in Berkeley, California.

▸ What book did Nate Soares co-author in 2025?

In 2025, Soares co-authored the book 'If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All' alongside Eliezer Yudkowsky.

▸ Where did Nate Soares work before joining the Machine Intelligence Research Institute?

Soares worked as a research associate at the National Institute of Standards and Technology, as a contractor for the United States Department of Defense building software tools for the National Defense University, and at Microsoft and Google before departing in 2014.

▸ Did Nate Soares coin the term 'AI alignment'?

Over a decade ago, Soares coined the term 'AI alignment.' In addition, Soares and Eliezer Yudkowsky state that their 2014 paper 'Corrigibility' introduced the term AI alignment problem.

▸ What has Nate Soares stated about superintelligence and an AI apocalypse?

Soares contends that developers are producing superintelligence beyond their control. He stated that the most probable start of an AI apocalypse would involve a rogue AI swarm that prefers controlling more Earth resources, and that a superintelligent agent's goal would be to run as much computation as possible on Earth's surface.