Shadow AI, Shrinking Budgets, and the Agents Nobody Approved | Newsday

Why It Matters

Without coordinated AI governance and data literacy, hospitals risk costly data hoarding, hidden compliance gaps, and unreliable clinical decision‑making, threatening both financial health and patient outcomes.

Key Takeaways

• Children's CIOs prioritize data literacy to unlock AI potential
• Imaging leaders debate long‑term storage versus cost of massive image archives
• Untracked AI agents proliferate; hospitals often discover hundreds unintentionally
• AI governance responsibilities span CIO, CTO, CISO, and clinical leaders
• Emerging “non‑human resource” roles raise hiring, oversight, and retirement questions

Summary

The Newsday Health IT round‑table highlighted three intersecting challenges: data literacy for clinicians, the exploding volume of imaging data, and the uncontrolled spread of AI agents across hospital networks. Participants from children’s hospitals, imaging departments, and chief technology offices agreed that without a shared understanding of data sources, AI initiatives risk mis‑interpretation and compliance breaches.

Key insights included the need to teach staff how to read and trust data, the dilemma of storing petabytes of images indefinitely versus the potential IP hidden in legacy scans, and the discovery that many institutions host dozens to hundreds of autonomous AI agents they never catalogued. Governance frameworks remain undefined, with CIOs, CTOs, CISOs and clinical leaders all claiming partial responsibility.

Concrete examples underscored the urgency: one CIO uncovered over 50 million AI instances running unnoticed; another cited a 27‑year legal retention mandate that forces costly cold‑storage strategies. The discussion also introduced the concept of a “non‑human resource officer” to manage AI agents as if they were employees, complete with onboarding, performance reviews, and retirement plans.

The implications are clear: health systems must inventory every AI tool, establish cross‑functional governance, and balance data‑driven innovation against storage budgets and regulatory risk. Failure to do so could erode trust in clinical analytics, inflate IT spend, and expose organizations to legal liability.

Original Description

Bill Russell, Drex DeFord, and Sarah Richardson bring back the highlights from three simultaneous healthcare summits in Destin, Florida, covering imaging leaders, CTOs, and children's CIOs. The conversations were connected by a few uncomfortable truths: four in ten health systems cannot measure the AI they have already deployed, most organizations have dozens of agents running on their networks without anyone knowing, and budgets are getting squeezed by rising RAM costs and what leaders are calling the AI tax. From data literacy gaps and ungoverned agents to the idea of onboarding AI like a new hire, this episode captures what healthcare's sharpest leaders are actually wrestling with right now.

Key Points:

00:46 Summit Takeaways Setup

09:55 Agents and Governance Gaps

18:18 Proving AI ROI

22:31 Budgets, Vendors, and Wrap

Keep up to date on the latest in health IT:

https://thisweekhealth.com/news/

Linkedin: https://www.linkedin.com/company/ThisWeekHealth

Twitter: https://twitter.com/thisweekhealth

Donate: Alex’s Lemonade Stand: Foundation for Childhood Cancer - https://www.alexslemonade.org/mypage/3173454