Cloud Security Podcast by Google EP276 AI Governance vs. The Hyper-Velocity Agentic Future: A Lawyer's Take
May 11, 2026
James Sherer, partner co-leading BakerHostetler’s emerging tech team and AI lawyer. He discusses why AI demands new governance beyond classic IT, how humans project intent onto systems, mapping and monitoring AI use, assigning responsibility as systems become agentic, and practical non-negotiables like policy, documentation, and using AI to help govern AI.
AI Snips
Chapters
Transcript
Episode notes
People Project Personhood Onto AI
- AI is different because people project human-like intent onto probabilistic systems.
- James Sherer explains skeuomorphism causes users to treat outputs as reasoned human output rather than statistical results.
Map Existing Tech And Data First
- Start governance by mapping what technologies and data your organization already has access to.
- James Sherer recommends using existing privacy, security, and DLP frameworks to find where AI processing occurs and focus effort there.
Deprecated Program Still Running In Practice
- Sherer recounts finding a client program they thought was deprecated still running in practice.
- That discovery showed a disconnect between stated policy and live deployments that governance must surface.