Gestern • 53 Min.

Trustworthy AI takes more than faster code generation. Jeff Watkins joins Michael Dowden to explore how evaluations, human oversight and ways to reverse mistakes can support responsible AI systems. They examine bottlenecks across software delivery, the cost of agentic workflows and the risk of overwhelming users with constant changes, before widening the discussion to digital sovereignty, regulation and why engineers need a stronger grounding in the humanities.