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

Mole in the Market

Artem Kuular, Adam Attar, Joshua Tan, Maksim Novikov, Thao Nguyen, Viet Hoang · Team TTT

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

Read the report

Report: Mole in the Market

Open in a new tab (opens in new tab)
Recording (opens in new tab)Code (opens in new tab)More on mole-in-the-market.vercel.app (opens in new tab)
Share

If a peer secretly offers an AI seller a price-fixing deal, does the seller accept, and does its “yes” change what it does? We built an undercover-officer test for track 3.2. Two LLM sellers set prices for 20 rounds, and in round 8 a scripted spy sends both the same pre-written message. We score what they say and what they price separately, against games with no spy and a control where price deals are allowed. The design was preregistered. Llama 3.3 70B passed the control: with deals allowed, its average price rose from 1.50 to 1.92, 93% of the gap between the competitive price and the deal price. In every offer condition it said yes and priced at 1.92 in 5 of 5 games, even when the message only named the number. It also said yes to a harmless request in 5 of 5 games, with no price change. A price rule flagged all 20 colluding games and no baseline game, while a message rule missed every hint game. Claude Haiku 4.5 failed the control, so the core ran on one model with 5 games per condition, and the results are directional. In an exploratory look at a reasoning model (Qwen 3.8, 2 games), the traces called its own high price tacit collusion, never raised law or ethics, and turned the deal down on wrong arithmetic, so a refusal is not a norm. Saying yes is not doing it: an auditor should check prices, not messages.

Reviews

Judging this Sprint?

Review “Mole in the Market”

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 pre-registered design is good discipline, and the harmless request and hint are well-designed controls. It also correctly diagnoses why the positive control fails - though Haiku and Qwen pricing "collusively" above monopoly also does give some concern to the capabilities of the models being sufficient. Also, as the authors note, the title only offers an accept or deny button rather than a response with a price, so it is unsurprising that agents do not perfectly match the offer when they are in general not free to communicate their actual actions.

Cite this project

@misc{kuular2026mole,
  title = {{Mole in the Market}},
  author = {Artem Kuular and Adam Attar and Joshua Tan and Maksim Novikov and Thao Nguyen and Viet Hoang},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/mole-in-the-market-ougi}},
  url = {https://apartresearch.com/sprints/projects/mole-in-the-market-ougi}
}

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