BBC的分析文章聚焦于OpenAI在9月声称用88小时解决了纳维-斯托克斯问题，以及与纽约大学数学家特里斯坦·巴克马斯特和Anthropic的莱文特·阿尔波格之间的争议，争议点在于Codex存储的草稿是否影响了OpenAI。OpenAI否认曾获取这些草稿。
OpenAI Maths Dispute Highlights Fears AI Is Eroding Trust in Science Published Oct 28, 2026, the BBC analysis centers on OpenAI's September claim to have solved the Navier-Stokes problem in 88 hours and a dispute with NYU mathematician Tristan Buckmaster and Anthropic's Levent Alpoge over whether Codex-stored drafts influenced OpenAI. OpenAI denied access.
文章进而扩展到AI在科学领域的应用，提及Basecamp Research在伦敦的DNA收集、DeepMind的AlphaFold蛋白质结构预测突破及有争议的材料发现，以及关于AI生成论文、虚假引用和AI篡改同行评审的研究，还有Elsevier、Nature和arXiv等出版商采取的防范措施。
The article widens to AI in science, citing Basecamp Research's DNA collection in London, DeepMind's AlphaFold protein breakthrough and contested materials discovery, plus studies on AI-generated papers, fake citations and AI-altered peer review, and publisher safeguards from Elsevier, Nature and arXiv.