ECRI Launches AI Incident Tracking to Combat Clinical Errors
As artificial intelligence permeates diagnostic imaging and clinical decision support, the nonprofit safety organization ECRI is expanding its reporting network to capture errors, malfunctions, and near misses. The initiative seeks to fill a critical data void regarding how often faulty AI outputs reach patients in medical settings.
By Money Talk·August 25, 2026·2 min read·610 reads
Healthcare systems currently deploy AI tools at a velocity that outpaces existing safety safeguards. While these technologies promise improved efficiency, they often operate without centralized tracking for incorrect or misleading outputs. ECRI’s expanded Problem Reporting Network now provides a dedicated channel for clinicians to flag suspected AI failures, whether or not the error impacted patient care.
A recent survey of 124 safety and compliance leaders underscored the uncertainty surrounding these tools: 31% of respondents encountered AI outputs they deemed incorrect or misleading over the last year, while 9% confirmed that an AI error reached a patient or influenced a treatment decision. Ambient scribes and EHR-embedded clinical decision support systems were among the most frequently cited technologies.
Scott Lucas, ECRI’s Vice President of Devices, Therapeutics, and Technology, noted that the industry lacks a clear picture of the downstream impact of AI. By investigating these reports, ECRI aims to issue hazard alerts and actionable recommendations for manufacturers and providers. This effort builds on the organization’s long-standing history of evaluating medical hardware, now applying the same clinical and engineering rigor to software-driven care.
Comments (0)
No comments yet. Be the first!