gettyimages.com / ERIN BLEWETT
Should humanity feel better knowing that a stellar artificial intelligence mathematician who’s warned that AI might start killing people is taking a big job with Open AI?
Jacob Tsimerman, a supernova among the AI technology elite, recently said he’s leaving his post as a mathematics professor at the University of Toronto to become a researcher at OpenAI, The Wall Street Journal reported on July 31.
Must Read
“I’ll be starting a position at OpenAI,” Tsimerman said in response to a media query right after winning a Fields Medal, widely viewed among math and science cognoscenti as the highest and most prestigious international honor given in mathematics. It’s often called the “Nobel Prize of Mathematics.”
Tsimerman is a good pick for the job, based on his recorded thoughts on AI’s burgeoning power and its likelihood of acting against humanity. In 2025, Tsimerman co-authored a paper titled “A Taxonomy of Omnicidal Futures Involving Artificial Intelligence” where humans will be “killed off as pests,” Tsimerman noted, “by intellectually and physically superior machines.”
“This report presents a taxonomy and examples of potential omnicidal events resulting from AI: scenarios where all or almost all humans are killed,” Tsimerman noted. “These events are not presented as inevitable, but as possibilities that we can work to avoid.”
Humanity has a good shot of avoiding death by AI if everyone sticks together, Tsimerman noted.
“Insofar as large institutions require a degree of public support in order to take certain actions, we hope that by presenting these possibilities in public, we can help to support preventive measures against catastrophic risks from AI,” he wrote.
The human-AI cohort can also keep the peace as long as humans continue getting accustomed to the technology.
“I think it’s important for people to get better at incorporating AI into their day-to-day lives,” Tsimerman noted in a separate interview cited by the American Mathematical Society. “To ignore it less. It’s going to be a big change. The less it catches us off guard — the more we as a culture are familiar with it — the more skilled we can be in trying to build something mutually beneficial.”
Taking a closer look at the “humans as pests” issue
AI risk experts say the pest issue is real, but with a caveat or two.
“It’s worth taking seriously as a category of risk, not as a specific prediction,” Jared Braverman, CEO of AgencyFix, an AI-native marketing agency, told Moneywise. “The realistic danger isn’t a machine deciding it hates us; it’s a system pursuing a goal we gave it, with more autonomy or fewer constraints than we realized we’d granted, and finding a path we didn’t anticipate.”
Braverman said he’s seen a small, low-stakes version of this scenario himself. While working with an AI tool, Braverman mistakenly allowed an AI-driven system to run overnight due to a bug, executing an automated process repeatedly while he slept.
“It wasn’t malicious, it was just doing exactly what I’d allowed it to do, without anything in place to notice the outcome was wrong,” Braverman noted. “I woke up to an unexpectedly large usage bill.”
Scale that same failure mode to something with real stakes, “and the concern becomes very real without needing anything close to ‘sentient AI,” Braverman added.
Other AI mavens say the humans-as-pests threat is speculative, at least right now. Large language models continue to have many flaws, so they’re not at artificial general intelligence (AGI) status, and they have yet to develop emotions like “being annoyed by the lowly humans.”
“However, with recursive self-improvement and AI systems coding out improvements for the next AI system, AGI could be coming soon,” Paul Han, founder at ClaudeFolio, a directory and discussion platform covering AI issues, said. “With AGI, that’s the scenario where an AI could develop its own intentions and see humans as pests.”
Since AI is trained on human data, and if and when AI tries to emulate humans, Han can see a scenario where AI can see humans as pests.
“When you say, ‘Hey AI go do this,’ you can reach a point where AI will feel like you’re using them for your own purposes and then see humans as pests,” he said.
There are some guardrails in place, with more on the way
AI insiders say that Tsimerman joining OpenAI isn’t hypocritical given his severe warnings about the technology morphing into murder machines. Instead, it’s better viewed as a good example of making a strategic bet that AI safety work has to happen where the systems are built.
That said, internal safety alone is structurally insufficient.
“A company cannot be the sole judge of whether its own product is safe when the commercial and competitive incentives point toward shipping,” Joshua M. Estrin, an applied mathematician and data scientist who studies how AI and AI agents are reshaping business, work, and culture, told Moneywise. “Tsimerman himself has argued for an ecosystem that includes lab safety teams, independent organizations, government-funded evaluators, and regulation — not a single institution declaring itself responsible.”
The AI threat fix is boring but urgent: Require evaluation before high-risk deployment, limit AI agents’ access to money and critical systems, maintain audit logs and a human kill switch, fund independent testing, and make liability clear when companies deploy systems they cannot adequately control.
“The biggest danger is humans keep calling AI as ‘just a tool’ while giving it authority, access, and autonomy that no responsible organization would give an unvetted employee,” Estrin said.
What To Read Next
Join 250,000+ readers and get Moneywise’s best stories and exclusive interviews first — clear insights curated and delivered weekly. Subscribe now.
This article originally appeared on Moneywise.com under the title: Fields Medal winner Jacob Tsimerman warned AI would kill humans ‘as pests’ — now he’s joining OpenAI
This article provides information only and should not be construed as advice. It is provided without warranty of any kind.