AI system solves open math problem and proves it rigorously — no human mathematicians needed

Researchers built an AI system that tackled an unsolved problem in algebra, worked through it, then translated its own reasoning into machine-checkable formal proof with almost no human help. This matters because math research currently requires humans to verify every step — if AI can generate proofs that machines can verify, the bottleneck shifts from 'is this right?' to 'did the AI find anything worth publishing?'

Until now, AI could suggest math solutions but humans had to check them. This system generates solutions and independently verifies them using formal logic tools, which means the verification step no longer requires a human expert to read through the reasoning. The constraint changes from expertise to computation. What this enables: if the informal reasoning part gets faster or better at exploring hard problems, the formal verification keeps pace automatically. What it doesn't solve yet: whether AI is actually better at finding new problems worth solving, or just better at proving ones humans already suspected.