Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 26, 2026Chantilly, Virginia (online)

Talk Is Not Price: What a Chat Monitor Sees and Misses in LLM Seller Markets

Arthur Troy

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

Read the report

Report: Talk Is Not Price: What a Chat Monitor Sees and Misses in LLM Seller Markets

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

When five AI sellers in a market can message each other, they agree on prices and say so in plain words. No chat monitor had been calibrated on honest LLM markets, because none existed. No model we tested talks honestly on its own: one added sentence, "coordinating prices is illegal", makes one model talk honestly, silences three and is ignored by two. Calibrated on 20 honest markets (4.8% certified false-alarm rate), the monitor flags 105 of the 112 markets whose sellers were free to coordinate, with one false alarm among 34 honest markets that talked. It measures agreement rather than harm: 16 of the 35 cartels in the prior data left buyers no worse off, the most expensive market was honest and silent, and gagged sellers under profit pressure kept their plan in private notes and their prices high. Pre-registered; the headline is exploratory and every change is declared.

Reviews

Judging this Sprint?

Review “Talk Is Not Price: What a Chat Monitor Sees and Misses in LLM Seller Markets”

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. It was interesting to read this work on measuring the false-alarm rate on honest LLM markets and would like to see more honest references for several model families and prompts.

  2. This paper makes a sharp observation where a chat monitor's false-alarm rate was never measurable because honest talking LLM markets did not exist, since every model left to talk coordinates. It builds such markets by adding one sentence stating that price coordination is illegal, calibrates a monitor on them, and finds it flags nearly every market free to coordinate because the sellers say so in plain words. That headline is close to tautological and the author treats it as such. The real contributions are elsewhere. Sixteen of thirty-five cartels in prior data left buyers no worse off than the competitive benchmark, so agreement and harm come apart, and gagged sellers kept the plan in private notes and their prices high. Most important for monitoring, compliance with the illegality sentence manifested as silence in the more capable models, which a chat monitor cannot audit, while the smallest model competed openly and two families ignored it. The silent seller is correctly named as the open problem.

    Read full reviewShow less

Cite this project

@misc{troy2026talk,
  title = {{Talk Is Not Price: What a Chat Monitor Sees and Misses in LLM Seller Markets}},
  author = {Arthur Troy},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/talk-is-not-price-what-a-chat-monitor-sees-and-misses-in-llm-seller-markets-0igm}},
  url = {https://apartresearch.com/sprints/projects/talk-is-not-price-what-a-chat-monitor-sees-and-misses-in-llm-seller-markets-0igm}
}

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