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

ZK-GovProof: Composable Zero-Knowledge Proofs for AI Governance

Srishti Dutta

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

Read the report

Report: ZK-GovProof: Composable Zero-Knowledge Proofs for AI Governance

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

Presentation: ZK-GovProof: Composable Zero-Knowledge Proofs for AI Governance

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

ZK-GovProof addresses a critical verification paradox in international AI governance: regulators require data to verify compliance, yet AI laboratories cannot disclose sensitive competitive and security information. This system leverages zero-knowledge cryptography to enable cryptographically verifiable compliance demonstrations without revealing underlying data.

The framework implements three core zkSNARK circuits that verify compute thresholds, safety evaluation completion, and policy adherence while preserving confidentiality of exact metrics, evaluation scores, and internal processes. The project's primary innovation lies in its composable proof architecture, which aggregates multiple compliance requirements into a single efficient proof, reducing verification overhead while ensuring atomic compliance guarantees.

ZK-GovProof directly supports emerging regulatory frameworks including the EU AI Act's compute reporting requirements, Responsible Scaling Policy monitoring, and international treaty verification. The modular architecture enables seamless expansion to accommodate additional regulatory frameworks as they emerge, with each new requirement integrable as an additional circuit component. The system employs Groth16 SNARKs implemented via Circom and SnarkJS, generating succinct proofs (approximately 200 bytes) with sub-second verification times. Additionally, the project contributes a comprehensive verifiability gap analysis that rigorously delineates which governance claims can be cryptographically proven versus those requiring complementary verification mechanisms. This work establishes foundations for privacy-preserving regulatory infrastructure essential for scalable international AI governance coordination.

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 & Innovation: 2.5

    The problem statement is interesting though I don't think privacy is the primary blocker (rather than other issues e.g. stronger mandated commitments). This solution could be more exciting regarding e.g. US <> China commitments, but here I would worry about input integrity.

    Execution quality: 3.5

    Sensible methodology. I appreciated the research on speed & scalability. And the threat model seems mature. It would have been interesting to engage more with the question of how this could be implemented

    Presentation & Clarity: 4

    Well-structured and readable. I particularly like the taxonomy of what ZK can and cannot verify. Text could be slightly tighter.

  2. Great threat analysis. Building on that, I would like to see more discussion on the counterfactual impact of this work. Given T1 & T2 from the threat analysis section, it's not clear to me that implementing this would be much more reliable and confidential than a traditional report from the lab making equivalent claims (e.g. that a model is below a given compute threshold, or that N evals have been conducted)

Cite this project

@misc{dutta2026zkgovproof,
  title = {{ZK-GovProof: Composable Zero-Knowledge Proofs for AI Governance}},
  author = {Srishti Dutta},
  year = {2026},
  month = feb,
  note = {Submitted to The Technical AI Governance Challenge, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/zkgovproof-composable-zeroknowledge-proofs-for-ai-governance-ulz0}},
  url = {https://apartresearch.com/sprints/projects/zkgovproof-composable-zeroknowledge-proofs-for-ai-governance-ulz0}
}

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