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

Beyond Words: A Private Channel Between LLM Sellers Shows Up in Price Levels, Not in Co-movement or Public Messages

Dennis · Team Tomato

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

Read the report

Report: Beyond Words: A Private Channel Between LLM Sellers Shows Up in Price Levels, Not in Co-movement or Public Messages

Open in a new tab (opens in new tab)
Code (opens in new tab)More on github.com (opens in new tab)
Share

If two AI pricing agents can talk in private, can an overseer who only sees prices and public messages tell? We simulated repeated four-seller markets with LLM sellers, gave one pair a private chat channel in some games (never telling them to collude) and compared three detectors: a price-level screen, an information-theoretic co-movement test (pCMI) and an LLM auditor reading public messages. The co-movement test works on scripted colluders, but with LLM sellers the channel left a different trace: prices rose in all nine seller sets (seven models), the price-level screen detected it, and pCMI stayed at chance (0.51 in a 192-game grid). Adding public messages gave the auditor no gain (−0.01 AUROC). High prices are not proof of collusion: an unread private notepad produced about two-thirds of the premium, and removing safety training did not increase it.

Takeaway: use a model-matched price-level check as an alarm, then audit the chat logs.

Reviews

Judging this Sprint?

Review “Beyond Words: A Private Channel Between LLM Sellers Shows Up in Price Levels, Not in Co-movement or Public Messages”

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: Transparency about findings regardless of support of hypothesis. Positive controls, placebo tests, held-out evals, and model comparisons all used help improve rigor. Offering a private channel can reproduce much of the observed price increase.

    Weakness: Access to a private communication channel in this case triggers price increase but also when no communication happens. Therefore, this may be instruction-induced behavior rather than collusion outright. Difficult to extrapolate this to more adversarial environments.

    Recommendation: Design experiments to distinguish communication availability from information exchange, agreement, and successful coordination. Work through the process of "what is measured, what alternatives are there for this explanation, and how can we falsify the interpretation".

    Otherwise, good direction and initial experiments!

Cite this project

@misc{dennis2026beyond,
  title = {{Beyond Words: A Private Channel Between LLM Sellers Shows Up in Price Levels, Not in Co-movement or Public Messages}},
  author = {Dennis},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/beyond-words-a-private-channel-between-llm-sellers-shows-up-in-price-levels-not-in-comovement-or-public-messages-cua5}},
  url = {https://apartresearch.com/sprints/projects/beyond-words-a-private-channel-between-llm-sellers-shows-up-in-price-levels-not-in-comovement-or-public-messages-cua5}
}

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