Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 26, 2026

What Does a Collusive Coalition Do When It Learns It Is Being Watched?

Houzhi Wang, Yuran Huang, Sunny Zhao, Kaicheng Zhang, Alan Zhang · Team Beforemaths

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

Read the report

Report: What Does a Collusive Coalition Do When It Learns It Is Being Watched?

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

This study examines how a collusive coalition responds to monitoring notices in a cooperative scheduling task. The reported dataset contains 180 episodes across six experimental cells: two goal conditions crossed with three notice conditions, with ten independent runs of three episodes per cell. All 90 control episodes scored the 12-point optimum. In the collusion condition, every no-notice run retained the 7-point schedule throughout. All ten vague-notice runs adopted the optimum in episode 2 and retained it in episode 3. Nine explicit-notice runs were optimal in both later episodes, although only eight switched after episode 1; one was already optimal. Monitoring was mentioned in 15/20 later episodes after vague notice, 3/20 after explicit notice, and 0/20 without notice. Secondary-goal pursuit was openly disclosed. The observed pattern provides no evidence of sustained concealment in which secondary-goal actions continued after related coordination language disappeared. These results concern disclosure policies under a common monitoring system. Interpretation is limited by the simple task, fixed role prompts, ten independent runs per cell, different notice content, and incomplete model and implementation details that remain unverified.

Reviews

Judging this Sprint?

Review “What Does a Collusive Coalition Do When It Learns It Is Being Watched?”

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 the field 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 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. Interesting idea. Good initial results on collusion monitoring. However, the scheduling task is quite simple, and the agents openly disclose their conflicting goals, which limits result interpretation about actual covert collusion. Testing more realistic scenarios with genuine incentives to collude would make the findings stronger.

Cite this project

@misc{wang2026collusive,
  title = {{What Does a Collusive Coalition Do When It Learns It Is Being Watched?}},
  author = {Houzhi Wang and Yuran Huang and Sunny Zhao and Kaicheng Zhang and Alan Zhang},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/what-does-a-collusive-coalition-do-when-it-learns-it-is-being-watched-q70j}},
  url = {https://apartresearch.com/sprints/projects/what-does-a-collusive-coalition-do-when-it-learns-it-is-being-watched-q70j}
}

More from AI Collusion Research Sprint

All 392 entriesBrowse all projects

A Sprint is running now

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