When an AI Agent Goes Rogue, Who's Responsible? | Civo Navigate London

Civo

Oct 13, 2026

AI agents are taking real actions in production. So when one misbehaves, who's accountable: the model provider or the company running it?

David Sullivan explains why a regulated bank can't outsource accountability to an AI lab. James Faure shares how Clairo AI uses hypothesis testing and quality gates before any agent reaches production. Dr Rosemary Francis argues that guardrails which amount to a strongly worded letter won't survive. What agents need are genuine sandboxes and a limited blast radius.

Watch the full AI panel: https://youtu.be/P3srrj_KdOs
====================
A new era of cloud computing, where speed, simplicity and predictable billing unite. Civo makes complexity a thing of the past.
Sign up to Civo -► https://www.civo.com

Subscribe to our YouTube Channel -► http://civo.io/subscribe

Follow Civo: