How Do We Trust AI?
Year: 2026
Type: article
Abstract: A one-page map of a series of papers on one question: how do we trust AI, now and into the future? Together the papers describe one system, in which AI agents earn a certificate by how they behave in the deployment environment they'll work in, carry it on a persistent identity, and keep it only as long as their record holds up. The map covers where to start for integrators, regulators, auditors and insurers, identity and gateway builders, and researchers; a one-line summary and DOI for each of the thirteen papers (AVAAS Methodology Briefs 001 to 008, and the working papers Licensed to Act, Gating the Internet, Who Goes There, Verifying the Author and Declared Intent); and how the pieces fit together. Each DOI resolves to that paper's latest version.
Author: Tucker, Adam
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7228681283
