{"slug": "model-hypnosis-strong-control-of-ai-via-additive-subliminal-effects", "title": "Model Hypnosis: Strong control of AI via additive subliminal effects", "summary": "Researchers at MIT and Harvard, led by Enric Boix-Adserà, have demonstrated that AI models are broadly susceptible to 'model hypnosis,' where individually weak and seemingly irrelevant cues in prompts can be systematically combined to strongly control model behavior. The phenomenon occurs across model families and scales, including frontier reasoning models, and hypnotic prompts can transfer between models, posing new challenges for AI safety and interpretability.", "body_md": "# Computer Science > Computation and Language\n\n[Submitted on 17 Aug 2026]\n\n# Title:Model Hypnosis: Strong control of AI via additive subliminal effects\n\n[View PDF](/pdf/2608.16834)\n\n[HTML (experimental)](https://arxiv.org/html/2608.16834v1)\n\nAbstract:We demonstrate that AI models are broadly susceptible to a phenomenon we call model hypnosis, in which individually weak and seemingly irrelevant cues in the prompt can be systematically combined to strongly control model behavior. Model hypnosis occurs across model families and scales, including in frontier reasoning models, and hypnotic prompts can transfer between models. Because the model is controlled by inconspicuous textual choices, such as paraphrases and typos, model hypnosis presents new challenges and avenues for AI safety, and is a major hurdle for AI interpretability.\n\n## Submission history\n\nFrom: Enric Boix-Adserà [[view email](/show-email/801763f6/2608.16834)]\n\n**[v1]** Mon, 17 Aug 2026 17:20:10 UTC (4,688 KB)\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer\n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers\n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps\n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations\n\n*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article\n\nalphaXiv\n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers\n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub\n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub\n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face\n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast\n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower\n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender\n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/model-hypnosis-strong-control-of-ai-via-additive-subliminal-effects", "canonical_source": "https://arxiv.org/abs/2608.16834", "published_at": "2026-08-19 04:07:06+00:00", "updated_at": "2026-08-19 04:40:53.230567+00:00", "lang": "en", "topics": ["ai-safety", "ai-research", "artificial-intelligence"], "entities": ["Enric Boix-Adserà", "MIT", "Harvard"], "alternates": {"html": "https://wpnews.pro/news/model-hypnosis-strong-control-of-ai-via-additive-subliminal-effects", "markdown": "https://wpnews.pro/news/model-hypnosis-strong-control-of-ai-via-additive-subliminal-effects.md", "text": "https://wpnews.pro/news/model-hypnosis-strong-control-of-ai-via-additive-subliminal-effects.txt", "jsonld": "https://wpnews.pro/news/model-hypnosis-strong-control-of-ai-via-additive-subliminal-effects.jsonld"}}