Can AI identify reliable home BP monitors? New study raises concerns

Clinical takeaway: Do not rely solely on AI chatbots to determine whether a home blood pressure monitor has been clinically validated. Instead, verify devices using established registries, such as ValidateBP.org, before recommending them to patients.

Home blood pressure monitoring plays a key role in hypertension diagnosis and management, but inaccurate devices can lead to inappropriate clinical decisions. Although guidelines recommend validated monitors, patients and clinicians increasingly turn to AI tools for quick answers about device reliability. New research presented at the American Heart Association (AHA) Hypertension Scientific Sessions 2026 suggests these tools may be unreliable for this purpose; the findings are preliminary and have not yet been peer reviewed.

Researchers evaluated four popular AI tools—ChatGPT, Microsoft Copilot, Perplexity, and Google Gemini—using 324 home blood pressure monitors, including 145 validated and 179 unvalidated devices. Validation status was assessed using established registries, including STRIDE BP, ValidateBP, and Hypertension Canada.

Each AI tool was asked three differently worded questions about each device's validation status, ranging from a general consumer-style query to a specific request referencing an official registry.

Google Gemini performed best, correctly classifying devices 86% to 91% of the time, depending on question wording. ChatGPT, Copilot, and Perplexity were less reliable, with accuracy ranging from 63% to 83%.

All four tools were less accurate at identifying validated monitors than unvalidated ones, sometimes incorrectly rejecting devices listed on official registries. Results were also inconsistent: When researchers repeated queries for a subset of devices that had produced mixed results, answers frequently changed across testing sessions.

The findings highlight a potential limitation of AI-assisted clinical decision-making, particularly when patients use chatbot responses to select home monitoring equipment. Performance may change as AI technology evolves, but researchers emphasize the importance of independently verifying device validation.

“We found that most AI tools performed only slightly better than if you had flipped a coin for each question,” said presenting author Anna Soriano, MD, of the University of Montreal. “Even Google Gemini, which performed best, was often wrong and couldn’t find information that is easily located.”

Source: Soriano A. (2026 Oct 8) American Heart Association Hypertension Scientific Sessions 2026. Abstract TH125. Accuracy of artificial intelligence tools in identifying validated home blood pressure monitors