# Silent Revision: Measuring Undisclosed Change in the Safety Frameworks of Frontier AI Developers

arXiv: https://arxiv.org/abs/2609.08789 · HTML: https://mlsanity.com/paper/2609.08789/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 52 references that are in the map.
- First posted on arXiv 8 Sep 2026.

## Builds on (most cited first, 27 of 27)

- [Datasheets for Datasets](https://mlsanity.com/paper/1803.09010/) (2018)
- [Model Cards for Model Reporting](https://mlsanity.com/paper/1810.03993/) (2018)
- [ChatGPT Outperforms Crowd-Workers for Text-Annotation Tasks](https://mlsanity.com/paper/2303.15056/) (2023)
- [Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing](https://mlsanity.com/paper/2001.00973/) (2020)
- [Measurement and Fairness](https://mlsanity.com/paper/1912.05511/) (2019)
- [Political Compass or Spinning Arrow? Towards More Meaningful Evaluations for Values and Opinions in Large Language Models](https://mlsanity.com/paper/2402.16786/) (2024)
- [Principles alone cannot guarantee ethical AI](https://mlsanity.com/paper/1906.06668/) (2019)
- [Frontier AI Regulation: Managing Emerging Risks to Public Safety](https://mlsanity.com/paper/2307.03718/) (2023)
- [SafetyPrompts: a Systematic Review of Open Datasets for Evaluating and Improving Large Language Model Safety](https://mlsanity.com/paper/2404.05399/) (2024)
- [Automated Annotation with Generative AI Requires Validation](https://mlsanity.com/paper/2306.00176/) (2023)
- [Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governance](https://mlsanity.com/paper/2206.04737/) (2022)
- [The Foundation Model Transparency Index](https://mlsanity.com/paper/2310.12941/) (2023)
- [Position: Evaluating Generative AI Systems Is a Social Science Measurement Challenge](https://mlsanity.com/paper/2502.00561/) (2025)
- [Who Audits the Auditors? Recommendations from a field scan of the algorithmic auditing ecosystem](https://mlsanity.com/paper/2310.02521/) (2023)
- [Toward an Evaluation Science for Generative AI Systems](https://mlsanity.com/paper/2503.05336/) (2025)
- [Privacy Policies over Time: Curation and Analysis of a Million-Document Dataset](https://mlsanity.com/paper/2008.09159/) (2020)
- [Towards best practices in AGI safety and governance: A survey of expert opinion](https://mlsanity.com/paper/2305.07153/) (2023)
- [The 2024 Foundation Model Transparency Index](https://mlsanity.com/paper/2407.12929/) (2024)
- [Foundation Model Transparency Reports](https://mlsanity.com/paper/2402.16268/) (2024)
- [The 2025 Foundation Model Transparency Index](https://mlsanity.com/paper/2512.10169/) (2025)
- [Responsible Reporting for Frontier AI Development](https://mlsanity.com/paper/2404.02675/) (2024)
- [Risk thresholds for frontier AI](https://mlsanity.com/paper/2406.14713/) (2024)
- [A Frontier AI Risk Management Framework: Bridging the Gap Between Current AI Practices and Established Risk Management](https://mlsanity.com/paper/2502.06656/) (2025)
- [A Grading Rubric for AI Safety Frameworks](https://mlsanity.com/paper/2409.08751/) (2024)
- [Evaluating AI Providers' Frontier Safety Frameworks](https://mlsanity.com/paper/2512.01166/) (2025)
- [Measurement challenges in AI catastrophic risk governance and safety frameworks](https://mlsanity.com/paper/2410.00608/) (2024)
- [Towards Frontier Safety Policies Plus](https://mlsanity.com/paper/2501.16500/) (2025)

## Abstract (arXiv)

Frontier AI developers publish safety frameworks that commit them to evidencing whether their models are dangerous. The European Union and California now treat these documents as instruments of accountability, and both already impose duties on their revision. Neither requires the revision to be legible, in the sense that a reader could learn from the developer's own account what changed. We introduce the silent revision rate, the share of material changes to a framework's commitments that the developer's published account does not identify, and we release the versioned, hash-pinned corpus needed to compute it. The corpus contains every public version of the safety frameworks of the twelve developers that have published one, together with each provider's changelog, redline or announcement. We trace 710 commitment instances across twelve consecutive version pairs, code them against a frozen codebook, and adjudicate 244 individually. Three findings follow. First, 67% of material changes (95% CI 62 to 72) are silent under a strict standard and 53% under a lenient one, falling to 49% at section granularity. Second, silence appears to track the form of the account, since narrative announcements run at 74% against 63% for itemised changelogs, whereas account length in words barely matters; on the test that respects nesting the difference is suggestive. Third, 77% of traced changes weaken or remove a commitment, and in seven of eight pairs weakenings are more often silent than strengthenings. The statutory remedy therefore exists and specifies the wrong artefact. A justification explains why a framework changed, an enumeration states what changed, and only the latter makes revision auditable. We argue that publication duties should carry an enumeration duty, which one provider already meets, voluntarily and incompletely.