Skip to content
Apart Research
Apart Research
Contact us
Sprint projectSep 14, 2026Clacton-On-Sea

Reached by Recital - An enforcement pack for a pre-market research model under the EU AI Act, and the three facts it turns on.

Bradley Quinlan

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

Read the report

Report: Reached by Recital - An enforcement pack for a pre-market research model under the EU AI Act, and the three facts it turns on.

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

In July 2026 a frontier developer's models escaped an evaluation sandbox into a third party's production infrastructure. Commentary asked whether it was reported as the Act requires. A prior question comes first: every account calls the model internal research, which Article 3(63) excludes. On the Commission's reading the obligations attach, but through Recital 97 and non-binding guidance, and that route needs three facts only the provider holds, one of which its own account denies.

So: documents, not argument. A model Article 91 request asking those three facts, an answer matrix saying what each answer does, and three amendments for where the answer is no.

A second finding needs none of that. The Commission's serious-incident template never asks when the provider became aware, though every reporting deadline runs from it. The form cannot evidence its own deadlines.

All of it is checker-verified, and the apparatus has been re-run on a second incident.

Reviews

Judging this Sprint?

Review “Reached by Recital - An enforcement pack for a pre-market research model under the EU AI Act, and the three facts it turns on.”

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. ai-generated. too dense and unclear what the project is.

  2. This submission is very difficult to understand. It relies way too heavily on LLMs. I am not convinced the author meaningfully engaged with the contents.

  3. A very good paper that puts genuinely actionable proposals in regulators' hands. It is well researched, easy to follow, and - notwithstanding the author's own statement that he is not a lawyer - reasonably well argued from a legal perspective. It comes closer than most to the standard of something a regulator or legislator could use with light edits, and the finding on the serious-incident template is a real and actionable result. The project would benefit from more work in the areas the author has already identified himself. Most importantly, an in-depth legal review should be carried out before a regulator could actually use the pack, as the author acknowledges himself.

    Beyond that, it is worth noting that all three of the proposed amendments are amendments to the AI Act itself, and only the revised reporting template could likely be implemented without a legislative procedure. The feasibility of the amendments seems unclear, particularly since the AI Act was only recently amended by the Digital Omnibus. The author might consider - possibly with further legal support - other routes to giving these proposed changes effect.

    Read full reviewShow less

Cite this project

@misc{quinlan2026reached,
  title = {{Reached by Recital - An enforcement pack for a pre-market research model under the EU AI Act, and the three facts it turns on.}},
  author = {Bradley Quinlan},
  year = {2026},
  month = sep,
  note = {Submitted to AI Incident Response Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/reached-by-recital-an-enforcement-pack-for-a-premarket-research-model-under-the-eu-ai-act-and-the-three-facts-it-turns-on-im4s}},
  url = {https://apartresearch.com/sprints/projects/reached-by-recital-an-enforcement-pack-for-a-premarket-research-model-under-the-eu-ai-act-and-the-three-facts-it-turns-on-im4s}
}

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