Patient safety nonprofit ECRI is expanding its Problem Reporting Network to capture and investigate errors, malfunctions, and near misses involving AI tools and AI-enabled devices used in patient care.

ECRI is calling on healthcare providers, health systems, and clinicians nationwide to submit reports of any incident where an AI-enabled tool may have contributed to an error or introduced risk into care delivery. ECRI triages and investigates those reports and shares findings with each submitter. This program compliments the ECRI and ISMP Patient Safety Organization (PSO), which has gathered and analyzes more than 8 million reports of safety events from healthcare providers nationwide.

The organization says that clinical AI tools and AI-enabled medical devices are being deployed at a pace that has challenged the safeguards designed to prevent, identify, and respond to failures.

AI now assists with everything from diagnostic imaging and clinical decision support to personal chatbots, but there is no centralized, healthcare-specific mechanism tracking how often these tools produce incorrect or misleading outputs, and how often those outputs reach a patient.

“ECRI has persistently emphasized the risk of adopting AI with insufficient scrutiny,” said Scott Lucas, Ph.D., ECRI’s vice president of devices, therapeutics, and technology, in a statement. “Although we appreciate AI’s tremendous potential, we don’t yet have a clear picture of its downstream impact in healthcare. Without a robust reporting dataset and analysis, the industry cannot sufficiently improve the design and integration of AI tools and devices. We must look to evidence and data to understand the evolving risks and associated system factors, to enable the use of the safest, most effective technologies.”

In a recent ECRI survey of 124 respondents — mostly quality, safety, risk or compliance leaders — nearly one-third said they encountered an AI output they believed was incorrect or misleading over the past year (31%), while 34% had not, and 35% were unsure. Nine percent reported an AI error reached a patient or impacted a care decision. Ambient scribes were the most commonly encountered AI tools in the survey (37%), ahead of EHR-embedded clinical decision support (31%) and clinical LLM assistants or chatbots (31%).

ECRI has evaluated the effectiveness and clinical evidence surrounding numerous AI-enabled devices and AI tools, including applications in imaging, fall prevention, anesthesia, wearable technology for chronic disease management, cardiovascular diagnoses, colonoscopy systems, and behavioral therapy.

ECRI's Problem Reporting Network offers a free, confidential channel for reporting medical device and technology problems. Every submission is triaged and investigated by ECRI's clinical and engineering experts, and ECRI follows up directly with the submitter to share what was found.

When a safety risk is identified in a device or technology, ECRI informs relevant stakeholders including manufacturers, providers, and regulatory agencies by issuing hazard reports that describe the problem and provide actionable recommendations to reduce risk. The Problem Reporting Network also contributes to the top health technology hazards and patient safety concerns in research reports ECRI publishes for the public each year.