Skip to content
Apart Research
Apart Research
Contact us
Sprint projectJan 11, 2026Madison, WI

Even the Best AI Would Hurt Us

Chris Santos-Lang · Team MAD Chairs

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

Read the report

Report: Even the Best AI Would Hurt Us

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

MAD Chairs may be the first work of game theory which combines the real-world significance of the Prisoner’s Dilemma with chess’s defiance of human mastery. This study predicts the consequences of adding AI players to real-world manifestations of MAD Chairs, such as crowded traffic, the limited attention of social media, large representative government, and scarce real estate. oTree code, as typically used for behavioral economics, is open sourced on GitHub, facilitating both reproducibility and extension to human trials, but the subjects for this study take the form of AI which approximate behavior previously observed in human subjects, as well as the current grandmaster strategy and strategies suggested by Gemini, ChatGPT, Claude, DeepSeek and Qwen. The results indicate that adding even the best-behaved AI possible to our ecosystem would hurt us in real-life MAD Chairs situations unless we place trust in machines as one must now do to maintain grandmaster status in chess.

Reviews

Judging this Sprint?

Review this project

Your public critique appears on this page without your name. Your private critique is not published; only the Apart team reads it. If you agree below, we share your review with grantmaking.ai (opens in new tab) and the Transformative AI Fund so strong projects can be funded.

Not shown on this page.

Shown on this page, without your name.

Only the Apart team reads this, and funders if you agree below.

Share my name publicly on grantmaking.ai *
Share my private critique with funders *

How much would this matter for AI safety 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 AI safety 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.

  1. Very interesting theoretical work. Thought-provoking and a fresh angle in the ai-safety space. I do feel that this used as a proxy for human behaviour is bit reductive, as it does not capture how humans can strategically adapt, unite in demanding situations to survive. Empirical experiments based on agents to simulate this could yield more interesting ideas.

  2. 1. There is a clear tournament design, but philosophical conclusions are somewhat an overreach for what the data supports.

    2. The core argument is fine and understandable, but overall the length of writing is bloated.

    3. Could talk more about the "So what?" question, and potential impacts of what to do with this framework.

Cite this project

@misc{santoslang2026even,
  title = {{Even the Best AI Would Hurt Us}},
  author = {Chris Santos-Lang},
  year = {2026},
  month = jan,
  note = {Submitted to AI Manipulation Hackathon, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/even-the-best-ai-would-hurt-us-ji74}},
  url = {https://apartresearch.com/sprints/projects/even-the-best-ai-would-hurt-us-ji74}
}

More from AI Manipulation Hackathon

All 70 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026