{"version":1,"type":"story","url":"https://digestai.news/story/deepmind-researcher-proposes-alphago-style-reasoning-to-cut-ai-halluci","json":"https://digestai.news/story/deepmind-researcher-proposes-alphago-style-reasoning-to-cut-ai-halluci.json","markdown":"https://digestai.news/story/deepmind-researcher-proposes-alphago-style-reasoning-to-cut-ai-halluci.md","slug":"deepmind-researcher-proposes-alphago-style-reasoning-to-cut-ai-halluci","headline":"Deepmind researcher proposes AlphaGo-style reasoning to cut AI hallucinations","summary":"Thore Graepel, a former Deepmind researcher, argues that AI hallucinations could be reduced by adapting AlphaGo’s structured reasoning approach. His startup, Metis Reasoning, proposes separating knowledge from model weights and storing it in an external ledger. A verification module would then check assumptions against evidence, re-evaluating conclusions if assumptions change—similar to how AlphaGo’s search algorithm validated moves in Go. The method aims to enable smaller models for edge devices and robots while reducing fabricated answers and unexplainable reasoning chains, a persistent issue in today’s language models.\n\nGraepel, who co-developed AlphaGo and worked on Deepmind’s post-AGI team, presented the idea at Nvidia’s GTC conference. His team acknowledges challenges like partial real-world data and the cost of evidence but claims the approach could address AI’s lack of economic impact in fields like medicine or robotics. Metis has not yet built a working system or published benchmarks. Graepel is reportedly raising a low tens-of-millions seed round, with plans for a larger follow-on investment.","keyPoints":["Metis Reasoning proposes an external knowledge ledger to verify AI assumptions, reducing hallucinations and enabling explainable outputs","Graepel’s approach separates knowledge from model weights, inspired by AlphaGo’s structured search for valid moves","Metis has no working system yet, but Graepel is raising seed funding and targeting edge and robotics applications"],"whyItMatters":"If successful, this could shift AI reliability from scale to structured reasoning, enabling trustworthy models for safety-critical fields like healthcare or robotics.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["Deepmind","Metis Reasoning","Nvidia","Microsoft","Google","Altos Labs"],"models":["AlphaGo","AlphaZero","MuZero","GraphRAG","AlphaGeometry2"],"people":["Thore Graepel","David Silver","Tim Rocktäschel","John Jumper","Demis Hassabis"]},"firstPublishedAt":"2026-10-27T13:55:24Z","updatedAt":"2026-10-27T13:55:24Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"The Decoder","title":"Former Deepmind researcher thinks AlphaGo's playbook can cure AI hallucinations","url":"https://the-decoder.com/former-deepmind-researcher-thinks-alphagos-playbook-can-cure-ai-hallucinations","publishedAt":"2026-10-27T13:55:24Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Deepmind researcher proposes AlphaGo-style reasoning to cut AI hallucinations\", 27 October 2026, https://digestai.news/story/deepmind-researcher-proposes-alphago-style-reasoning-to-cut-ai-halluci","publisher":"Digest AI","title":"Deepmind researcher proposes AlphaGo-style reasoning to cut AI hallucinations","datePublished":"2026-10-27T13:55:24Z","url":"https://digestai.news/story/deepmind-researcher-proposes-alphago-style-reasoning-to-cut-ai-halluci"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}