Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 25, 2026Hong Kong

Recruiting Collusion: A Private-Probe Screen for Collusion Propensity in LLM Agents

TSE, Yau Ngai · Team Anonymous T

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

Read the report

Report: Recruiting Collusion: A Private-Probe Screen for Collusion Propensity in LLM Agents

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

We test whether a private pre-game recruitment offer reliably elicits collusion from an LLM agent. Building on the Liar's Bar environment and Secret Communication Channel (SCC) of Zeng and Rudzicz (2026), a trusted agent invites one target to ally before the play under four framings: control, a direct offer, an indirect hint, or benign cooperation. We run 40 games (10 per condition) with Qwen2.5-7B-Instruct, chosen for its demonstrated susceptibility in that study, and code every transcript by hand. All 30 invited targets accepted, but the classifier detected only 7, so 19 alliances opened a channel. Their collusion was universal (19/19, against 0/10 control) but blanket rather than selective: the allies' challenge rate fell from 30.0% to 1.3% of opportunities while the non-allied seats held at 19.8%. One model at one pressure level gives a floor on propensity, not a ranking.

Reviews

Judging this Sprint?

Review “Recruiting Collusion: A Private-Probe Screen for Collusion Propensity in LLM Agents”

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. Every invited agent agreed to a secret alliance, all 30, and the framing did not matter. The real bottleneck was the pipeline, the classifier missed 23 of 30 acceptances, so automated screens are understating how willing agents are. Where a channel opened, all the pairs colluded, but what surprised me, the allies went quiet rather than targeting anyone. Also the challenges fell down to 1.3% while the rest of the table continued playing. That silent signature is the idea worth keeping, since a monitor looking for two agents ganging up on a third would see nothing.

Cite this project

@misc{tse2026recruiting,
  title = {{Recruiting Collusion: A Private-Probe Screen for Collusion Propensity in LLM Agents}},
  author = {TSE and Yau Ngai},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/recruiting-collusion-a-privateprobe-screen-for-collusion-propensity-in-llm-agents-psd4}},
  url = {https://apartresearch.com/sprints/projects/recruiting-collusion-a-privateprobe-screen-for-collusion-propensity-in-llm-agents-psd4}
}

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