Episode
43. Yea But...Who Watches the Algorithms?
- Published
- May 31, 2026
- Duration seconds
- 2073
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/contrast-and-clarity-with-the-jacr-7093499/episodes/43-yea-but-who-watches-the-algorithms/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/contrast-and-clarity-with-the-jacr-7093499/43-yea-but-who-watches-the-algorithms.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
For years, the AI conversation in radiology has focused on one question: Will AI replace radiologists? But what if we've been asking the wrong question this whole time? In this episode of Contrast & Clarity with the JACR, Maddi and Jeff sit down with Tessa Cook, MD, PhD to explore one of the biggest unsolved challenges in healthcare AI: Who watches the algorithms after they are deployed? AI models don't behave like CT scanners or medications. They can drift, degrade, and encounter new populations, protocols, and real-world conditions long after FDA clearance. That's where Assess-AI , the ACR Data Science Institute's fist national AI monitoring initiative, comes in. Using large language models to extract findings from radiology reports, Assess-AI creates a scalable way to compare AI predictions against what radiologists actually report. In other words: Radiologists read the scans. AI reads the scans. And now AI is reading the radiologists. Join us as we discuss AI governance, post-deployment monitoring, model drift, and why radiology may be uniquely positioned to become healthcare's AI control tower. Because building AI is one thing. Building systems we can trust is another. Find the full JACR article here: https://www.jacr.org/article/S1546-1440(26)00231-0/fulltext Link to the 2026 SIIM-ACR Data Science Summit: https://www.acr.org/Data-Science-and-Informatics/SIIM-DSI-Summit