Send us Fan Mail

This episode explains how health systems can govern the shadow AI healthcare staff already use, starting with the controls they have today instead of waiting for perfect ones.

Aaron Bach is co-founder and CTO of Liminal, with more than 15 years across software, AI and enterprise architecture. He joins Chris Hutchins, Founder and CEO of Hutchins Data Strategy Consultants, to argue that AI should be governed at the behavior layer while it runs. Their conversation moves from unsanctioned tools to model choice, human review and accountability when something goes wrong.

Aaron shares early research Liminal ran in 2023, in which about three out of four employees said they already used AI at work. Roughly two thirds of that group admitted using generative AI tools their organization had not approved, mostly to keep up with heavy workloads. He explains why the NIST AI RMF places this work in its Manage function, which calls for ongoing action over a policy sitting on a shelf. Chris and Aaron also discuss why the best use cases often come from front office staff and clinicians, and why committee-heavy planning tends to stall.

Blocking and shaming users tends to push AI use further out of sight, so governance has to serve both security teams and the people doing the work. No single platform covers every AI need, since AI spans roughly 15 to 20 distinct product categories. Model output can look polished while being subtly wrong, which makes trained human reviewers more critical than ever. Unreviewed AI output sent straight to patients, such as an incorrect portal message about a cancer diagnosis, shows where review must sit.

Aaron frames active AI oversight through the Manage function of the NIST AI Risk Management Framework. He describes behavioral automation, an approach that observes how individual employees work to surface use cases specific to them. A multi-model strategy reflects his view that different models suit different problems. Claude Cowork comes up as an example of a tool built on a coding harness, well suited to some tasks and poorly matched to others.

Aaron Bach co-founded Liminal in 2023 and serves as its CTO. Before Liminal, he led engineering at Fortune 500 scale, served as CTO of a pharmacy savings platform and ran engineering innovation at FIS. His work focuses on helping organizations in healthcare and other regulated industries put AI to practical use with governance in place.

Episode page: https://signalroompodcast.com/episodes/shadow-ai-governance-healthcare
Liminal: https://liminal.ai/
Aaron Bach on LinkedIn: https://www.linkedin.com/in/aaronbach
Beneath the Signal, Chris's book on the human work behind trusted data and responsible AI in healthcare: Get it on Amazon
The Signal Room is featured in Feedspot's Best AI in Healthcare Podcasts, 2026: https://podcast.feedspot.com/ai_health_podcasts/

Support the show

About The Signal Room: The Signal Room is a podcast and communications platform exploring leadership, ethics, and innovation in healthcare and artificial intelligence. Hosted by Christopher Hutchins, Founder and CEO of Hutchins Data Strategy Consultants. Leadership, ethics, and innovation, amplified.

Website: https://www.hutchinsdatastrategy.com

LinkedIn: https://www.linkedin.com/in/chutchins-healthcare/

YouTube: https://www.youtube.com/@ChrisHutchinsAi

Book Chris to speak: https://www.chrisjhutchins.com