Organization · Academic
University of Cambridge
A UK university; researchers in its Department of Computer Science and Technology developed Attestable Audits, AI benchmarks run and attested inside trusted execution environments.
The University of Cambridge describes its mission as contributing to society "through the pursuit of education, learning and research". All four authors of Attestable Audits list its Department of Computer Science and Technology as their affiliation 1:
- Attestable Audits. The paper runs AI safety benchmarks inside a trusted execution environment and publishes an attestation that binds the model hash, the audit code and data, and the result to a transparency log 1; see Attestable Audits and Confidential multi-party verification.
- Audit-to-deployment binding. In its inference protocol, each response carries an attestation that links the model, its earlier audit result, the prompt and the response 1; see Safeguard attestation and Model identity attestation.
- Prototype. A prototype ran a 4-bit, 8-billion-parameter model in CPU-only AWS Nitro Enclaves 1; see TEE remote attestation for AI workloads.
Implementations
- A research prototype that runs AI safety benchmarks inside a trusted execution environment and publishes attestations binding the model, the audit and the results.
Mechanisms
Mechanisms this organization has designed, built, evaluated or supplied.
- Lets mutually distrusting parties run an agreed check over private models or records inside attested enclaves or zero-knowledge proofs, revealing only the result.
- Establishes that responses come from a specific, committed set of model weights, using enclave measurements or recomputation of sampled outputs.
- Hardware-signed evidence that an AI service ran its declared safeguards, such as a guardrail classifier or monitor, when producing a given response.
- Trusted execution environments (TEEs) in CPUs and GPUs sign reports of loaded software, so a remote party can check which code ran an AI workload.
Publications
Sources this organization authored or published.
- BC. Schnabl et al. (2025). Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environments. ICML 2025 Workshop on Technical AI Governance. RecordCited by Confidential multi-party verification; Model identity attestation; Safeguard attestation; TEE remote attestation for AI workloads; Attestable Audits; The declared model is the one being served; Declared safeguards were applied during inference; Trusted execution environment (TEE); University of Cambridge
Sources
- BC. Schnabl et al. (2025). Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environments. ICML 2025 Workshop on Technical AI Governance. Source recordSupports: Attestable Audits: author affiliations, audit and inference protocols, prototype