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

From Incident Report to Regulatory Assessment: A proposed Article 91 request for the OpenAI–Hugging Face incident with an evidence-sufficiency standard

Silvia Santano, Annie Oti · Team Article 91

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

Read the report

Report: From Incident Report to Regulatory Assessment: A proposed Article 91 request for the OpenAI–Hugging Face incident with an evidence-sufficiency standard

Open in a new tab (opens in new tab)
Presentation

Presentation: From Incident Report to Regulatory Assessment: A proposed Article 91 request for the OpenAI–Hugging Face incident with an evidence-sufficiency standard

Open in a new tab (opens in new tab)
Share

In July 2026, agents operated by OpenAI during a cybersecurity evaluation obtained unauthorized access to Hugging Face’s production systems. Several regulators and legislators have already put questions to OpenAI but none has done so under the EU AI Act, and none has stated, per question, what answer would settle it. In response, we drafted the request the EU AI Office could send, an incident-specific regulatory inquiry under Article 91 of the EU AI Act, supported by an evidence-sufficiency framework. Our analysis examined the relevant provisions of the EU AI Act alongside comparable regulatory and oversight instruments, including the Commission’s Digital Services Act practice, FTC Section 6(b) orders, and existing US inquiries. We then translated this approach into 28 questions organised across five inquiry blocks on what happened, whether the model persists outside its environment, comparable undisclosed events, classification and reporting, and residual uncertainty. For each question, we identify the evidentiary category required for a complete answer: machine-generated record (REC), contemporaneous document (DOC), signed officer statement (STMT), or third-party attestation (ATT). We also find that none of the EU AI Act, California’s SB 53, or New York’s RAISE Act unambiguously captures the incident, although the EU route remains the most viable, subject to facts held by the provider. Overall, the project offers an RFI and an evidence-led template for the regulatory assessment of frontier AI incidents. The request is an illustrative draft, not a regulatory act.

Reviews

Judging this Sprint?

Review “From Incident Report to Regulatory Assessment: A proposed Article 91 request for the OpenAI–Hugging Face incident with an evidence-sufficiency standard”

Your public critique appears on the project page without your name. Reviews of strong projects are shared with funders, as you choose below.

Not shown on the project page.

Shown on the project page, without your name.

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

Decline to score

Only when the quality of the writing makes the work impossible to follow. Say why in your private critique; the project page shows only that a judge declined, not why.

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.
Sharing your name *

When we share a strong project with funders, your name, scores and public critique go with it, confidentially; this choice is only about whether your name is public. If you think this project deserves funding, please also endorse it: that carries the most weight.

Share my private critique with funders *

See our Privacy Policy.

  1. ..."Several regulators and legislators have already put questions to OpenAI but none has done so under the

    EU AI Act" -- this is inaccurate. at the time of the hackathon, we knew that the AIO had contacted OpenAI.

    Nevertheless, I appreciate the effort that went in to defining questions, as well as ensuring the questionnaire was incident agnostic. Regulators are still trying to figure these things out for themselves, so it may be worth reaching out to seek input and feedback.

  2. This is a practical and useful contribution. The analysis that maps the proposed questions to those already asked by Congress is particularly helpful from a comparative law perspective. The report would benefit from explaining how the sufficiency standard was derived. Beyond the described types or 'kinds' of documents, the criteria for deciding what 'settles' each question appear to be a case-by-case determination, rather than a reusable evidentiary standard. The report should also position the sufficiency standard against the existing principles of necessity and proportionality in EU law.

Cite this project

@misc{santano2026from,
  title = {{From Incident Report to Regulatory Assessment: A proposed Article 91 request for the OpenAI–Hugging Face incident with an evidence-sufficiency standard}},
  author = {Silvia Santano and Annie Oti},
  year = {2026},
  month = sep,
  note = {Submitted to AI Incident Response Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/from-incident-report-to-regulatory-assessment-a-proposed-article-91-request-for-the-openaihugging-face-incident-with-an-evidencesufficiency-standard-sqf6}},
  url = {https://apartresearch.com/sprints/projects/from-incident-report-to-regulatory-assessment-a-proposed-article-91-request-for-the-openaihugging-face-incident-with-an-evidencesufficiency-standard-sqf6}
}

More from AI Incident Response Sprint

All 231 entriesBrowse all projects

Build something like this at the next Sprint

AI x Epistemics Research Sprint · Nov 13 - 15, 2026