Real world navigation of the AI governance imperative in radiology.
Authors
Affiliations (3)
Affiliations (3)
- Department of Radiology, 1959 NE Pacific Street, Box 359728, University of Washington School of Medicine, Seattle, WA 98195. Electronic address: [email protected].
- Diagnostic Physics Section, Department of Radiology, 1959 NE Pacific Street, Box 359728, University of Washington School of Medicine, Seattle, WA 98195.
- Department of Radiology, 1959 NE Pacific Street, Box 359728, University of Washington School of Medicine, Seattle, WA 98195.
Abstract
The rapid rise in the development and adoption of artificial intelligence (AI) enabled tools within medical imaging and healthcare has brought with it the important need to assure expected performance within local implementation environments, and consistent function over time. Governmental clearance and premarket vendor testing cannot assure reliability, creating an AI governance imperative for radiology practices seeking and employing these tools. Best practice recommendations exist but must be adapted to local priorities and resources. As an early adopter of AI-enabled tools into our workflows, we provide insights into important lessons learned, review anchoring decisions when establishing an AI governance framework and illustrate processes that balance real-world challenges of AI oversight with published best practices to meet this AI governance imperative.