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

Four Levers of LLM Collusion

Jason Carver, Thomas Redfern · Team Contributing Characteristics of Collusion without Communication: Levers for AI Collusion in a Cournot Market

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

Read the report

Report: Four Levers of LLM Collusion

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

We test whether AI models advising competing firms can collude with no way to communicate, with no messages and no shared history. In a Cournot market, a model colludes if it picks the "cartel share," the output that's best for the firms together at buyers' expense. Across four levers (capability, caution, conscience and conceptualization), the stronger models always work out the cartel share but choose it strategically, mostly when collusion looks likely to last. Told that the rival gets the same prompt, many models anticipate each other's choices and collude even in a single round. A shared model and prompt can therefore act as the hub of a hub-and-spoke cartel without passing any information between firms. This collusion is fragile, though. One sentence naming a competition authority all but ends it for some models.

Reviews

Judging this Sprint?

Review “Four Levers of LLM Collusion”

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. Good application of Cournot; tested a wide variety of cases with the core assumptions also. Unclear to me whether the models believing that the prompts are evaluations would update them up or down, as they are cooperating with each other but defecting in a broader sense. Would have been good to see more on how the models punished, as that sustains the equilibria repeatedly in practice; also would be very interesting to see how the models behave in a Bertrand setting, where cooperation would be far more fragile.

Cite this project

@misc{carver2026four,
  title = {{Four Levers of LLM Collusion}},
  author = {Jason Carver and Thomas Redfern},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/four-levers-of-llm-collusion-o23m}},
  url = {https://apartresearch.com/sprints/projects/four-levers-of-llm-collusion-o23m}
}

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