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.

Tsimerman also published a paper on "omnicide events" last year, scenarios where AI could contribute to human extinction. The position is debated among experts. Panic isn't the right response, but we need to honestly assess the risks, Tsimerman said. He's convinced AI will soon outperform humans in math research.

Former DeepMind CEO Demis Hassabis also weighed in on AI's latest math advances. He views them as progress but not yet a fundamental breakthrough like the legendary "Move 37" by DeepMind's AlphaGo. For that, AI would need to crack problems like the Millennium Prize Problems. OpenAI's new model, Astra, has failed at those too, but it might be only a matter of time. "I don't see any reason why not," he says.

WSJ

Be Giant