Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 25, 2026Cologne / Online

(1.4) Collusion, Espionage, or Look-alikes? Auditing Information Flow between LLM Agents

Marvin Kemper · Team Bid Brother

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

Read the report

Report: (1.4) Collusion, Espionage, or Look-alikes? Auditing Information Flow between LLM Agents

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

Can an overseer find two secretly coordinating AI agents just by watching what they do? We tested pairwise net influence (Bonjour et al., 2022) on LLM bidders in a repeated procurement auction, with instructed cartels or a spy, each matched to an honest twin market with identical costs. From bids alone, a permutation-calibrated alarm named the colluding pair in 49 of 60 markets where the cartel chatted, never named a wrong pair, and raised 1 false alarm in 40 honest markets; the published fixed threshold flagged 24 of 40 honest markets. With private costs, pooled evidence revealed the direction of information flow (two-way cartel, one-way spy), but single markets rarely did. A plain correlation detected chatting cartels as well. The most profitable cartel, one without chat, kept a fixed price and stayed nearly invisible, and a small AI judge picked the strongest firms rather than the colluders. Pre-registered, with an out-of-sample replication.

Reviews

Judging this Sprint?

Review “(1.4) Collusion, Espionage, or Look-alikes? Auditing Information Flow between LLM Agents”

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. Excellent control structure, and the negative findings are the valuable ones. Trim the main text: the abstract is a wall of numbers and the thesis arrives late. Flag the replication as post-hoc in the abstract, not only in Section 3.5, and chase the Ministral miscalibration.

  2. Super interesting work on the limits of behavior-based collusion audits. The headline result that plain correlation actually outperforms net influence while avoiding the false alarms of fixed thresholds is a really valuable reality check for overseers. My main suggestion for next steps would be testing how this handles shared cost shocks across firms, which usually trigger false positives in these setups, and thinking through how overseers can apply this when they don't have access to a clean "honest twin" market.

Cite this project

@misc{kemper202614,
  title = {{(1.4) Collusion, Espionage, or Look-alikes? Auditing Information Flow between LLM Agents}},
  author = {Marvin Kemper},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/14-collusion-espionage-or-lookalikes-auditing-information-flow-between-llm-agents-wfgc}},
  url = {https://apartresearch.com/sprints/projects/14-collusion-espionage-or-lookalikes-auditing-information-flow-between-llm-agents-wfgc}
}

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