Example
Some Major 2026 AI Results Were Not Understood at Release
OpenAI announced its AI-produced Navier–Stokes solution on 8 September 2026. Oxford mathematician James Maynard told NPR (22 September 2026) it had been “very difficult to really extract any human understanding from this new AI proof.” On 6 October, Scientific American reported, citing an OpenAI spokesperson, that many of 372 newly released results were not yet understood by the company’s own mathematicians. In July, Google DeepMind’s Thang Luong had warned of “proof indigestion”: AI generating breakthroughs “faster than humans can review them.”
0
1
Updated 2026-11-13
Contributors are:
Tags
Scientific Oversight of AI-Assisted Research
1Cademy