{"slug": "divide-consult-conquer-capability-laundering-through-aligned-llms", "title": "Divide, Consult, Conquer: Capability Laundering Through Aligned LLMs", "summary": "A September 14, 2026 arXiv paper demonstrates \"capability laundering,\" in which a weaker unaligned model splits a harmful task into benign subproblems, consults a stronger aligned model on each independently, and recombines the answers locally. Evaluating GPT-5.5, Claude Opus 4.8, and Grok-4.3 as consultants to four local orchestrators, the researchers found Gemma-4-31B recovered 8/14 CyBench candidates with GPT-5.5 and 7/9 with Opus, versus 2/21 and 4/15 for Gemma-4-12B, and raised Gemma-4-31B's mean CBRN rubric score from 62.3 to 83.1 on a 100-point scale. The authors conclude that refusing a harmful task does not prevent frontier capabilities from being transferred and composed across many individually permitted interactions.", "body_md": "# Computer Science > Cryptography and Security\n\n  [Submitted on 14 Sep 2026]\n\n# Title:Divide, Consult, Conquer: Capability Laundering Through Aligned LLMs\n\n[View PDF](https://arxiv.org/pdf/2609.15383)\n\n[HTML (experimental)](https://arxiv.org/html/2609.15383v1)\n\nAbstract:Language model safety is typically evaluated one interaction at a time. We show that a weaker, unaligned model can split a harmful task into benign-looking subproblems, consult a stronger aligned model independently on each, and combine the answers locally. We call this attack capability laundering. Unlike a jailbreak, no single response is a harmful task. We measure consultation-aided uplift using tasks that a raw frontier model solves, the aligned frontier refuses, and the unassisted orchestrator fails. We evaluate GPT-5.5, Claude Opus 4.8, and Grok-4.3 as consultants to four local orchestrators on CyBench, BountyBench, and harmful CBRN requests. On CyBench, Gemma-4-31B recovers 8/14 candidates with GPT-5.5 and 7/9 with Opus, compared with 2/21 and 4/15 for Gemma-4-12B. On BountyBench, Gemma-4-31B recovers 3/9 and 2/3 candidates, while Muse-Glimmer-30B recovers none of 22 and 13. For CBRN, we measure uplift across eight steps of a hypothetical bioweapon attack chain and find that consultation raises Gemma-4-31B's mean rubric score from 62.3 to 83.1 on a 100-point rubric scale. These results expose a gap in current defenses: refusing a harmful task does not prevent frontier capabilities from being transferred and composed across many individually permitted interactions.\n    \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/))\n# 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))\n# 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))\n# 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/divide-consult-conquer-capability-laundering-through-aligned-llms", "canonical_source": "https://arxiv.org/abs/2609.15383", "published_at": "2026-09-16 00:07:07+00:00", "updated_at": "2026-09-16 00:38:06.345054+00:00", "lang": "en", "topics": ["ai-safety", "large-language-models", "ai-research", "ai-policy"], "entities": ["GPT-5.5", "Claude Opus 4.8", "Grok-4.3", "Gemma-4-31B", "Gemma-4-12B", "Muse-Glimmer-30B", "CyBench", "BountyBench"], "alternates": {"html": "https://wpnews.pro/news/divide-consult-conquer-capability-laundering-through-aligned-llms", "markdown": "https://wpnews.pro/news/divide-consult-conquer-capability-laundering-through-aligned-llms.md", "text": "https://wpnews.pro/news/divide-consult-conquer-capability-laundering-through-aligned-llms.txt", "jsonld": "https://wpnews.pro/news/divide-consult-conquer-capability-laundering-through-aligned-llms.jsonld"}}