AI assurance and accountability
What evidence about an AI system is adequate for a particular decision, held by a particular role, in a particular deployment?
Claims about an AI system should be checkable by someone who was not involved in building it.
Most assurance work asks whether a model is good. That is the wrong unit. A model is not deployed, a system is, into an institution, for a decision, under a regulation, with someone accountable for the outcome. The same model can be adequately evidenced for one of those situations and badly evidenced for the next.
My work here builds methods and software that make that distinction operational: analysing risk as a function of deployment context, reasoning over regulatory obligations as executable statements rather than prose, and testing whether a body of audit evidence actually supports the decision it is offered for, including the cases where plausible-looking evidence does not.
Selected work
- PaperBA-FedSHAP: A reproducible toolkit for auditing background-induced attribution driftAuthor, SSRN2026
- PolicyAI in health: how smart technology is breaking new frontiers in medical careAuthor, Digital Skills and Jobs Platform, European Commission2026
- ProjectAn assurance toolchain for deployed AIAuthor2026
- SoftwareCSA-Lite, Context-Sliced AI Assurance LiteAuthor2026
- SoftwareLEXON-Bench, Executable AI Regulatory Obligation ReasoningAuthor2026
- SoftwareRCEA Passport Engine, Role-Conditioned Evidentiary AdequacyAuthor2026
- SoftwareProvena, reproducibility package for evidence-laundering stress testsAuthor2026
- SoftwareBA-FedSHAP, auditing background-induced attribution driftAuthor2026
Research projects
An assurance toolchain for deployed AI
A connected set of reproducible tools for deployment-conditioned risk analysis, executable regulatory reasoning, evidence adequacy scoring and adversarial stress testing.
Next / Territory
Privacy-preserving collaborative AI