Skip to content
Apart Research
Apart Research
Contact us
Sprint projectSep 14, 2026Toronto

What Nobody Signed: An Article 91 Request for the Audit Trail Behind the 2026 OpenAI Containment Failures

Khushi Suresh Rana

Submitted to AI Incident Response Sprint. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: What Nobody Signed: An Article 91 Request for the Audit Trail Behind the 2026 OpenAI Containment Failures

Open in a new tab (opens in new tab)
More on github.com (opens in new tab)
Share

Two OpenAI containment failures are public: one disclosed in twelve days, one unreported for months. This paper drafts the Article 91 request the Commission would have to send, eleven questions each marked for whether the law clearly permits it. Neither the AI Act nor the Code of Practice OpenAI signed requires a name to be attached to the decision. It proposes named sign-off at three points, backed by unannounced inspection.

Reviews

Judging this Sprint?

Review this project

Your public critique appears on this page without your name. Your private critique is not published; only the Apart team reads it. If you agree below, we share your review with grantmaking.ai (opens in new tab) and the Transformative AI Fund so strong projects can be funded.

Not shown on this page.

Shown on this page, without your name.

Only the Apart team reads this, and funders if you agree below.

Share my name publicly on grantmaking.ai *
Share my private critique with funders *

How much would this matter for AI safety if it worked? How innovative is it? For scores of 4-5: is this actually new to the field, or replicating recent work?

Scoring guide
  1. 1Negligible. No clear problem addressed, or no meaningful novelty.
  2. 2Limited. Addresses a real problem but with a generic or well-trodden approach. Incremental at best.
  3. 3Moderate. Clear problem with a reasonable approach; some novelty in framing or method beyond routine application of existing tools.
  4. 4Significant. Important problem with an original approach, or identifies a neglected problem area. A valuable contribution others could build on.
  5. 5Exceptional. Tackles a critical AI safety problem with a genuinely novel approach, or opens a new research direction. Clear theory of change. You'd be excited to share this with researchers in the area.

How sound are methodology, implementation, and findings?

Scoring guide
  1. 1Seriously flawed. Methodology broken, results uninterpretable, or implementation doesn't work.
  2. 2Weak. Approach has significant gaps: missing validation, flawed experimental design, or incomplete implementation.
  3. 3Competent. Technically solid given the short duration. Methodology makes sense, results are interpretable, limitations acknowledged, work builds toward clear conclusions.
  4. 4Strong. Thorough methodology with convincing validation. Results clearly support conclusions. Immediately useful for future work.
  5. 5Exceptional. Ambitious scope executed rigorously. Surprising findings, novel methods, or unusually robust validation.

How clearly are work, findings, and impact potential communicated?

Scoring guide
  1. 1Incomprehensible. Cannot determine what the project is actually claiming or doing.
  2. 2Hard to follow. Key information buried, missing, or diluted by excessive length. Significant effort to extract main points.
  3. 3Clear enough. Can understand the problem, approach, and results without undue effort. Core content clearly present: problem, method, findings, limitations.
  4. 4Well presented. Easy to follow, well-structured, appropriate level of detail. Target audience would get it quickly.
  5. 5Exceptionally clear. A pleasure to read. Complex ideas made accessible. Could serve as a model for how to present this type of work.

  1. I would have liked to have seen more discussion of the questions you chose in the official paper itself. Overall, I like the questions you landed on, but I wonder if they could be more standardized into an "incident response questionnaire" that the AIO could send out to any company. Right now, the questions are tailored to openai, but in a world where these things happen a lot, this format requires the AIO to tailor a questionnaire to each specific incident, which may be too much for the AIO to handle.

  2. This is an excellent idea buried in far too much LLM-generated text. A 1-2 page explanation of the proposal and the context, written clearly, with a half page on why this is an EU Commission power, followed by a far shorter Appendix A (which didn't try to restate what it is first, and didn't try to explain inside of the request why it was allowed and debating what it was allowed to do) would have been far better.

  3. The paper outlines a novel adapted instrument for enforcement of the EU AI Act in light of the OpenAI / Hugging Face incident. The paper does well in diagnosing and addressing genuine gaps in the audit trail, and proposes an interesting reframing that focuses on named individual accountability rather than simply what information the EU Commission can demand. The evidence is generally handled well and the methodology is clear.

Cite this project

@misc{rana2026nobody,
  title = {{What Nobody Signed: An Article 91 Request for the Audit Trail Behind the 2026 OpenAI Containment Failures}},
  author = {Khushi Suresh Rana},
  year = {2026},
  month = sep,
  note = {Submitted to AI Incident Response Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/what-nobody-signed-an-article-91-request-for-the-audit-trail-behind-the-2026-openai-containment-failures-u7zv}},
  url = {https://apartresearch.com/sprints/projects/what-nobody-signed-an-article-91-request-for-the-audit-trail-behind-the-2026-openai-containment-failures-u7zv}
}

More from AI Incident Response Sprint

All 233 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026