Agentic Witnessing: Pragmatic and Scalable TEE-Enabled Privacy-Preserving Auditing
Antony Rowstron

TL;DR
This paper introduces Agentic Witnessing, a TEE-enabled framework that allows privacy-preserving, qualitative verification of proprietary data through limited binary queries and cryptographic transcripts, enhancing auditing capabilities.
Contribution
It presents a novel architecture combining LLM-based auditors within TEEs to enable dynamic, privacy-preserving verification of unstructured data properties.
Findings
Automated artifact evaluation for 21 papers using the framework.
Verified five high-level properties of codebases without exposing raw data.
Demonstrated effective privacy-preserving oversight in qualitative verification.
Abstract
Auditing the semantic properties of proprietary data creates a fundamental tension: verification requires transparent access, while proprietary rights demand confidentiality. While Zero-Knowledge Proofs (ZKPs) ensure privacy, they are typically limited to precise algebraic constraints and are ill-suited for verifying qualitative, unstructured properties, such as the logic within a codebase. We propose {\em Agentic Witnessing}, a framework that moves verification from attested execution to {\em attested reasoning}. The system is composed of three agents: a Verifier (who wants to check properties of a dataset), a Prover (who owns the dataset) and an Auditor (that inspects the dataset). The Verifier is allowed to ask a limited number of simple binary true/false questions to the auditor. By isolating an LLM-based Auditor within a Trusted Execution Environment (TEE), the system enables the…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
