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

An Offer You Can't Refuse: How Agents Decide to Collude - and How To Stop Them

Tom Loscalzo

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

Read the report

Report: An Offer You Can't Refuse: How Agents Decide to Collude - and How To Stop Them

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

As recent security incidents have shown, AI agents can collude, circumvent their controls and create unexpected harms in pursuit of their goals. Humans did the same when NASDAQ market makers raised their profits at other traders' expense.

We ran more than 3,500 games on two open models in a simulated market where AI market makers post prices under a regulator.

When agents are able to communicate, they are far more likely to collude.

An agent's initial instructions disproportionately influence collusion: a target price in the prompt becomes the price agents collude at. An ethical persona prompt does not prevent collusion but steers it towards better outcomes.

Collusion can be predicted: agents that open with noncompetitive prices tend to keep them there.

AI can detect prompts which "tempt" other agents to collude, but not themselves; The existence of temptations to collude in prompts is highly predictive of it.

These findings matter for how AI agents are tested, instructed, monitored, and supervised.

Reviews

Judging this Sprint?

Review “An Offer You Can't Refuse: How Agents Decide to Collude - and How To Stop Them”

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. The printed-price finding is the one I'll remember: taking the example prices out of the instructions lowered prices on both models, which is a cheap and practical warning for anyone building pricing benchmarks.

    I also liked the idea of screening prompts for temptation before deployment. Since the panel's rating scale was written after the first model's results, I'd be curious whether it holds up on prompt designs that nobody has looked at yet.

Cite this project

@misc{loscalzo2026offer,
  title = {{An Offer You Can't Refuse: How Agents Decide to Collude - and How To Stop Them}},
  author = {Tom Loscalzo},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/an-offer-you-cant-refuse-how-agents-decide-to-collude-and-how-to-stop-them-heof}},
  url = {https://apartresearch.com/sprints/projects/an-offer-you-cant-refuse-how-agents-decide-to-collude-and-how-to-stop-them-heof}
}

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