{"slug": "certified-in-theory-broken-in-practice-assumption-gaps-in-cryptographic-model", "title": "Certified in Theory, Broken in Practice: Assumption Gaps in Cryptographic Model", "summary": "A new study from researchers including Elisaweta Masserova reveals that cryptographic model certification (CMC) schemes built on zero knowledge proofs can be attacked by model providers who engineer training data, allowing a model to certify over 99% accuracy on an audit dataset but achieve less than 30% accuracy on fresh samples from the same distribution. The paper formalizes new security notions for CMC frameworks and proposes a generic protocol template to close this assumption gap.", "body_md": "# Computer Science > Cryptography and Security\n\n[Submitted on 23 Jul 2026]\n\n# Title:Certified in Theory, Broken in Practice: Assumption Gaps in Cryptographic Model Certification\n\n[View PDF](/pdf/2607.21839)\n\n[HTML (experimental)](https://arxiv.org/html/2607.21839v1)\n\nAbstract:Privacy-preserving machine learning auditing protocols allow auditors to assess models for properties such as accuracy or fairness, without revealing their internals or training data. This makes them especially attractive for auditing models deployed in sensitive domains such as healthcare or finance. For these protocols to be meaningful in real-world audit settings, though, their guarantees must reflect how the model will behave once deployed, rather than merely certifying its behavior during an audit. Existing security definitions often miss this mark: most certify model behavior only on a fixed audit dataset, without ensuring that the same guarantees generalize to other datasets drawn from the same distribution.\n\nAs we show, this gap allows a model provider to attack many cryptographic model certification (CMC) schemes built on secure zero knowledge proofs (ZKP) by carefully engineering training data, resulting in models that exhibit benign behavior during an audit, but pathological behavior in practice. For example, we empirically demonstrate that an attacker can certify that a model achieves over 99% accuracy on an audit dataset, but less than 30% accuracy on fresh samples from the same distribution.\n\nTo address this gap, we formalize rigorous cryptographic security notions tailored to CMC frameworks, introduce a generic protocol template, and prove that it satisfies these requirements. Our results thus offer both cautionary evidence about existing approaches and constructive guidance for designing secure, privacy-preserving ML auditing protocols.\n\n## Submission history\n\nFrom: Elisaweta Masserova [[view email](/show-email/f17f55d8/2607.21839)]\n\n**[v1]** Thu, 23 Jul 2026 22:06:12 UTC (1,159 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/certified-in-theory-broken-in-practice-assumption-gaps-in-cryptographic-model", "canonical_source": "https://arxiv.org/abs/2607.21839", "published_at": "2026-07-29 04:07:06+00:00", "updated_at": "2026-07-29 04:23:26.931735+00:00", "lang": "en", "topics": ["ai-safety", "ai-research", "machine-learning"], "entities": ["Elisaweta Masserova", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/certified-in-theory-broken-in-practice-assumption-gaps-in-cryptographic-model", "markdown": "https://wpnews.pro/news/certified-in-theory-broken-in-practice-assumption-gaps-in-cryptographic-model.md", "text": "https://wpnews.pro/news/certified-in-theory-broken-in-practice-assumption-gaps-in-cryptographic-model.txt", "jsonld": "https://wpnews.pro/news/certified-in-theory-broken-in-practice-assumption-gaps-in-cryptographic-model.jsonld"}}