Skip to content
Apart Research
Apart Research
Contact us
Sprint projectFeb 1, 2026London

AI Governance Transparency Ledger

AIGC · Team Zuzana Kapustikova

Submitted to The Technical AI Governance Challenge. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: AI Governance Transparency Ledger

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

Presentation: AI Governance Transparency Ledger

Open in a new tab (opens in new tab)
More on ai-governance-transparency-ledger-5lfyre3cwf6dtfpikclmma.streamlit.app (opens in new tab)
Share

A tamper-proof compliance verification system for frontier AI governance that enables labs, auditors, and regulators to coordinate on safety requirements without requiring full mutual trust.

Key Features: - Deployment Gate: Blocks AI model releases until all compliance requirements are met and safety concerns resolved

- Multi-Party Mirrors: Ledger replicated across lab, auditor, and government; any tampering is instantly detectable through hash comparison

- Anonymous Whistleblowing: Safety researchers can raise concerns with identity protection (real identities never enter the system)

- Cryptographic Integrity: Hash chains ensure that sycompliance records cannot be secretly modified

- Zero-Knowledge Proofs: Labs can prove compliance thresholds without revealing sensitive operational data

Tech Stack: Python, FastAPI, Streamlit, SHA-256 hash chains, Merkle trees

For demo guidelines, see: JUDGES_TESTING_GUIDE.md in the GitHub repository

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. Impact potential & innovation: 2

    The author identifies a vertification gap but is going too broad with the solution. It's unclear as to why hash chains, Merkle trees, zero-knowledge proofs, and the other proposed ideas are the right solution here. It would have been helpful to more clearly lay out the threat models to start, and then present solutions to address these.

    Execution quality: 3

    I appreciate the working software and the ambition to implement across multiple primitives (hash chains, Merkle trees, ZK proofs, whistleblower mechanisms). But breadth came at the cost of depth and I would have liked to see either deeper implementation of one primitive or more engagement with what real world integration would require

    Presentation & clarity: 2

    The paper is clear and concise, and the limitations section correctly identifies some key issues, such as labs committing false data initially. But again, it is lacking technical depth, which makes the points difficult to understand.

    Read full reviewShow less

  2. Cool attempt at implementing a piece of fundamental technical infrastructure we will need to have in place for international agreements. A few comments in the spirit of constructive criticism:

    - Are the zero-knowledge proof actually zero-knowledge? My understanding is that the proof here involves revealing the count and blinding factor to the verifier, and calling this ZK seems to be overclaiming.

    - The paper says the ledger is distributed and that it’s mirrored across key stakeholders, with tampering detected. But little seems to be said about how mirrors are synchronized, how conflicts are resolved, what happens during network partitions, or who is authoritative when mirrors disagree. What's described here seems closer to having multiple parties independently store copies and that they can compare hashes, which is useful but seems weaker than the claims suggest.

    - The cryptographic contribution is standard (hash chains, merkle trees, commitment schemes have rarely been applied to AI governance but they’re very well-known primitives). Would be interesting to see more work on developing a properly original protocol/construction/security proof.

    Read full reviewShow less

Cite this project

@misc{aigc2026ai,
  title = {{AI Governance Transparency Ledger}},
  author = {AIGC},
  year = {2026},
  month = feb,
  note = {Submitted to The Technical AI Governance Challenge, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/ai-governance-transparency-ledger-2149}},
  url = {https://apartresearch.com/sprints/projects/ai-governance-transparency-ledger-2149}
}

More from The Technical AI Governance Challenge

All 47 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026