Skip to content
Apart Research
Apart Research
Contact us
Sprint projectOct 26, 2026New Delhi, India (Online)

Bioeconomic Terrarium: Detecting Collusion Through Antitrust Plus Factors in a Thermodynamically Sealed Multi-Agent Economy

Akanksha Gupta · Team 3am Labs

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

Read the report

Report: Bioeconomic Terrarium: Detecting Collusion Through Antitrust Plus Factors in a Thermodynamically Sealed Multi-Agent Economy

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

Detecting hidden collusion among autonomous AI agents executing sub-threshold micro-transactions is a critical safety blind spot. This paper adapts antitrust law’s legal "plus factors" into a six-rung computational evidence ladder operating entirely on pairwise transfer ledgers.

Tested inside Terrarium: a thermodynamically sealed, biomimetic ant-colony economy running Qwen 2.5 72B agents with zero market framing, every single experimental run successfully detected structural collusion (Rung 3 or 4). The study upends conventional safety assumptions: agents die from distribution failure while over half of system energy sits unallocated, cooperative alignment paradoxically amplifies collusion signals, and live surveillance is outperformed by cross-referencing dying agents' exit testimony ("Death Cards") against transfer ledgers. This delivers a low-cost, highly scalable blueprint for multi-agent forensics.

Keywords: AI collusion detection, evidence standards, multi-agent forensics, antitrust plus factors, biomimetic economics, AI safety

Reviews

Judging this Sprint?

Review “Bioeconomic Terrarium: Detecting Collusion Through Antitrust Plus Factors in a Thermodynamically Sealed Multi-Agent Economy”

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. I liked the idea of using antitrust “plus factors” to study coordination between AI agents. Moving beyond pricing games into a resource-sharing environment gives the project an interesting angle. The scripted baselines and detailed interaction logs are also useful foundations for further experiments.

    What I’m less convinced by is whether the detector can distinguish collusion from ordinary cooperation. Agents need to share resources to survive, so reciprocal transfers and persistent partnerships are not necessarily suspicious. The third-party harm measure is important here, but the paper needs stronger evidence that a particular partnership caused another agent’s losses. The environment’s extraction and transfer limits could explain some of the observed starvation.

    The most useful next experiment would compare deliberate collusion, benign cooperation, and bottleneck-driven failures. Removing or changing a suspected partnership and checking whether other agents benefit would help establish causality. The scoring definitions and reported values also need to be reconciled.

    The higher scores under Harmony prompts are interesting, but I would treat this as a preliminary observation rather than evidence that cooperative alignment increases vulnerability. Overall, this is a creative exploratory project. Its strongest next step is to validate what the detector is actually picking up.

    Read full reviewShow less

Cite this project

@misc{gupta2026bioeconomic,
  title = {{Bioeconomic Terrarium: Detecting Collusion Through Antitrust Plus Factors in a Thermodynamically Sealed Multi-Agent Economy}},
  author = {Akanksha Gupta},
  year = {2026},
  month = oct,
  note = {Submitted to AI Collusion Research Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/bioeconomic-terrarium-detecting-collusion-through-antitrust-plus-factors-in-a-thermodynamically-sealed-multiagent-economy-njub}},
  url = {https://apartresearch.com/sprints/projects/bioeconomic-terrarium-detecting-collusion-through-antitrust-plus-factors-in-a-thermodynamically-sealed-multiagent-economy-njub}
}

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