On 24 June 2025 Neil Lawrence gave oral evidence to the House of Commons Treasury Committee as part of its inquiry into AI in financial services. He appeared alongside Sandra Wachter (University of Oxford) and Galina Andreeva (University of Edinburgh Business School) in a session covering questions Q93–130.

Neil framed the central risk as placing excessive confidence in AI — treating systems that predict and automate as if they carry judgment they do not have. He urged the Committee to prefer “honest brokers,” including independent scientists, over “issue advocates” with a financial interest in deployment. He also warned of systemic fragility, comparing poorly governed AI in markets to a slow-motion version of the 2010 flash crash, and argued that regulators need a change of mindset for technologies that evolve faster than static rulebooks.

Wachter pressed the same theme from the other direction: large language models are not “magic eight balls that always tell the truth,” but next-word predictors whose hallucinations and historical-data limits matter when firms use them in consequential decisions. Andreeva stressed independent challenge and validation of AI models, of the kind credit institutions already apply to internal risk models. Coverage of the session includes Compliance Corylated and Professional Adviser.

The appearance sits with the group’s AI Adoption programme: studying how institutions take up tools without mistaking supply of models for safe, accountable use.