mlsanityData to 14 Oct 2026

Privacy-Preserving AI Verification via Minimal Information Disclosure

arXiv 2608.02774 ↗First posted 3 Aug 2026cs.CR cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

AI verification crosses a trust boundary: a verifier must learn enough to establish an authorized claim, yet the same evidence can reveal sensitive details about the model, workload, or hardware. We introduce minimal information disclosure (MID), which designs and quantifies the information content of verifier-facing evidence itself. MID measures collateral leakage with conditional mutual information: what the release reveals about the protected property after the authorized result is known. MID is general by design: it can accommodate different verification goals, protected properties, evidence sources, and deployment constraints. To demonstrate MID's practicality, we evaluate it on four physical measurements and six verification tasks spanning execution type, hardware identity, compute scale, and model identity. These experiments use three mechanism-design variables--the evidence channel, collection policy, and release transformation--but MID is not limited to these choices and can accommodate other deployable mechanisms. Across these tasks, MID produces three releases with perfect held-out verification and zero measured collateral leakage, while the remaining tasks yield explicit privacy--utility frontiers. MID also supports ZKP-certified releases: we demonstrate our proposed linear-projection mechanism using a Groth16 zk-SNARK.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 32 of its 57 references that are in the map. First posted on arXiv 3 Aug 2026.

Builds on

32 of 57 references.

References · 32

  1. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  2. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,281
  3. 2021Membership Inference Attacks From First Principlescited by 878
  4. 2020Systematic Evaluation of Privacy Risks of Machine Learning Modelscited by 292
  5. 2021Proof-of-Learning: Definitions and Practicecited by 84
  6. 2024Open Problems in Technical AI Governancecited by 49
  7. 2023Privacy Side Channels in Machine Learning Systemscited by 40
  8. 2023Tools for Verifying Neural Models' Training Datacited by 22
  9. 2026International AI Safety Report 2026cited by 22
  10. 2023What does it take to catch a Chinchilla? Verifying Rules on Large-Scale Neural Network Training via Compute Monitoringcited by 21
  11. 2025TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inferencecited by 17
  12. 2024Verifiable evaluations of machine learning models using zkSNARKscited by 16
20 more
  1. 2024Trustless Audits without Revealing Data or Modelscited by 15
  2. 2022Proof-of-Learning is Currently More Broken Than You Thinkcited by 12
  3. 2025Verifying International Agreements on AI: Six Layers of Verification for Rules on Large-Scale AI Development and Deploymentcited by 11
  4. 2023DeepTheft: Stealing DNN Model Architectures through Power Side Channelcited by 10
  5. 2025Mechanisms to Verify International Agreements About AI Developmentcited by 9
  6. 2021Can one hear the shape of a neural network?: Snooping the GPU via Magnetic Side Channelcited by 8
  7. 2025Hardware-Enabled Mechanisms for Verifying Responsible AI Developmentcited by 8
  8. 2025Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environmentscited by 8
  9. 2025Verifying LLM Inference to Detect Model Weight Exfiltrationcited by 6
  10. 2026PAL*M: Property Attestation for Large Generative Modelscited by 5
  11. 2017Privacy-Utility Tradeoffs under Constrained Data Release Mechanismscited by 4
  12. 2025Flexible Hardware-Enabled Guarantees for AI Computecited by 4
  13. 2026Timing and Memory Telemetry on GPUs for AI Governancecited by 4
  14. 2026Hardware-Level Governance of AI Compute: A Feasibility Taxonomy for Regulatory Compliance and Treaty Verificationcited by 4
  15. 2026Bit-Exact AI Inference Verification Without Performance Tradeoffscited by 4
  16. 2026Detecting Hidden ML Training With Zero-Overhead Telemetrycited by 4
  17. 2026Agentic Witnessing: Pragmatic and Scalable TEE-Enabled Privacy-Preserving Auditingcited by 3
  18. 2025ShadowScope: GPU Monitoring and Validation via Composable Side Channel Signalscited by 2
  19. 2026Zero knowledge verification for frontier AI training is possiblecited by 2
  20. 2024Input-Dependent Power Usage in GPUscited by 1

Built on by

Not cited yet.