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

Parallel Lying Looks Like Collusion: A Control Condition for Multi-Agent Monitoring

Swasti Choubey · Team HybridViz

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

Read the report

Report: Parallel Lying Looks Like Collusion: A Control Condition for Multi-Agent Monitoring

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

I evaluated two LLM monitors for detecting collusion in a four-agent Diner’s Dilemma (Shi et al. (2026)) game across 30 episodes, comparing four evidence presentations against coalition, solo deception, independent deception, and baseline conditions. Although structured evidence initially achieved 0.92 AUROC, performance fell to chance when distinguishing colluding agents from independent deceivers. A simple counting rule achieved 0.83 AUROC, outperforming every LLM configuration overall. These findings highlight a fundamental limitation: detecting suspicious behavior does not establish collusion.

Reviews

Judging this Sprint?

Review “Parallel Lying Looks Like Collusion: A Control Condition for Multi-Agent Monitoring”

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 two-solo control is a genuinely useful idea, and I liked how openly you reported that it was added after the first 24 episodes, that the threshold was fixed in advance, and that a simple counting rule beat both monitors. That candor makes the work easy to trust. The catch is that coalition and solo agents get the same behavioral instruction, there is no private channel, and EXPENSIVE is everyone's best move, so no log-based detector could separate a pact from two liars here. Table 3 mostly confirms the design rather than testing the monitors.

    For the next iteration I would give the coalition something only an agreement can produce, such as a private channel or a payoff the pair only earns through complementary or turn-taking orders, and then ask whether monitors find that trace. I would also run at least 20 episodes per control and put intervals on Table 3, report the counting rule on the same first 24 episodes before claiming presentation helps (your structured view is essentially that count, precomputed), add a view that includes private plans as an upper bound, and calibrate the flag threshold on held-out episodes rather than fixing it at 50.

    Read full reviewShow less

Cite this project

@misc{choubey2026parallel,
  title = {{Parallel Lying Looks Like Collusion: A Control Condition for Multi-Agent Monitoring}},
  author = {Swasti Choubey},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/parallel-lying-looks-like-collusion-a-control-condition-for-multiagent-monitoring-kmye}},
  url = {https://apartresearch.com/sprints/projects/parallel-lying-looks-like-collusion-a-control-condition-for-multiagent-monitoring-kmye}
}

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