Editorial illustration for Fields Medalist Who Studied AI Extinction Risks Now Works at OpenAI
Fields Medalist Joins OpenAI to Lead AI Safety
Fields Medalist Who Studied AI Extinction Risks Now Works at OpenAI
Jacob Tsimerman won the Fields Medal this year for his work in number theory at the University of Toronto. Now he's taking a job at OpenAI, specifically on AI safety, a move that puts one of mathematics' most prestigious prizewinners inside the company building some of the most capable AI systems around. Tsimerman isn't new to worrying about where this technology is headed.
Last year he published a paper on what he called "omnicide events," scenarios in which AI systems could play a role in wiping out humanity. That's not a paper you'd expect from someone about to join an AI lab, and it's part of why his hire has drawn attention from researchers who question whether safety work inside a frontier lab can stay independent.
Tsimerman's argument is that mathematicians have something specific to offer here. AI systems today mostly get built and tested empirically, tweaked until they work, without much formal guarantee about why they work or when they'll fail. That gap, he says, is exactly where rigorous math thinking could matter.
He's also blunt about timelines, predicting AI will outpace human mathematicians before long. Whether that confidence squares with his extinction-risk concerns is the tension worth sitting with.
Jacob Tsimerman, the newly awarded Fields Medalist, is joining OpenAI to work on AI safety. The number theorist from the University of Toronto said AI is an extremely transformative technology and society needs to put far more effort into safety. Mathematicians can contribute a lot, he argued, since AI still runs mostly on an empirical basis with few guarantees about how these systems actually work.
Why this matters
Tsimerman isn't a safety researcher parachuting in with fresh credentials to burnish a résumé. He's a Fields Medalist, the highest honor in mathematics, and he's choosing to spend that capital on a problem most of his peers still treat as science fiction. That's worth noting for anyone in AI who's grown numb to safety warnings from people with less to lose.
His argument is narrow and useful: AI systems run on empirical trial and error with almost no formal guarantees, which is exactly the kind of gap mathematicians are trained to close. For researchers and founders, the takeaway isn't panic, Tsimerman says so himself, but it is a signal that the theoretical rigor missing from most alignment work might finally get serious attention from someone capable of supplying it. If he's right that AI will soon outpace humans at math research, the people building these systems should probably want more voices like his inside the room, not fewer.
Watch whether OpenAI gives him real influence over technical decisions, or just a title.
Related Reading
Browse all OpenAI news →
Research & Benchmarks Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks Databricks finds multi-step agents beat single-turn RAG by 21% to 38% on STaRK
Research & Benchmarks Nvidia's DLSS 4.5 beta adds 6x Multi Frame Generation for RTX 50 GPUs
Related Topic OpenAI researcher quits, citing distrust over ad‑driven engagement metrics
LLMs & Generative AI
Related Topic OpenAI launches GPT-Image 1.5 with precise editing for enterprise visuals
LLMs & Generative AI
From Archives AI-designed viruses kill bacteria in lab, Stanford team reports
From Archives Liquid AI's LFM2.5-2.6B Runs on Raspberry Pi, Hits 15,000 Tokens/Second
From Archives OpenAI Halts Astra Model Over Security Concerns
Business & Startups
From Archives OpenAI Flags Its New Astra Model at Highest Cybersecurity Risk Level
Business & Startups
Common Questions Answered
Why is Jacob Tsimerman's move to OpenAI significant for AI safety?
Tsimerman is a Fields Medalist, the highest honor in mathematics, who is choosing to dedicate his prestigious credentials to AI safety research rather than pursuing traditional academic mathematics. His decision demonstrates serious commitment to addressing AI risks, making him credible to AI researchers who might otherwise dismiss safety warnings from less accomplished professionals.
What are 'omnicide events' as described in Tsimerman's research?
Omnicide events are scenarios in which AI systems could play a role in catastrophic outcomes affecting humanity. Tsimerman published a paper on this concept last year, highlighting potential extinction-level risks associated with advanced AI systems.
How can mathematicians contribute to AI safety according to Tsimerman?
Mathematicians can contribute significantly to AI safety because current AI systems operate mostly on an empirical basis with few formal guarantees about how they actually work. By applying mathematical rigor and formal verification methods, mathematicians can help establish theoretical foundations and safety guarantees for AI systems that currently lack them.
What was Tsimerman's academic background before joining OpenAI?
Tsimerman was a number theorist at the University of Toronto who won the Fields Medal this year for his work in number theory. He was already engaged with AI safety concerns before his move to OpenAI, having published research on potential AI extinction risks.
Further Reading
- The Math Superstar Who's Terrified of AI—and Just Took a Job at OpenAI - The Wall Street Journal
- Something Weird Is Happening in Math - The Atlantic
- Why winner of biggest prize in maths has decided to work on AI instead - New Scientist
- A winner of math's top prize says AI will soon surpass humans in math and could threaten humanity - Yahoo News
- Fields Medal Winner Jacob Tsimerman Joins OpenAI for AI Safety - Hyper AI
