Turn on cookies to sign in
Signing in saves your progress to your Curie account. We can only do that with cookies on — turn them on to continue.
Don't trust AI to pick your blood pressure monitor. A new study found that popular AI tools — including ChatGPT, Copilot, and Gemini — frequently failed to correctly identify whether home blood pressure monitors were clinically validated. Even the best-performing tool, Google Gemini, was wrong up to 14% of the time. Experts say patients should consult their clinician or check official validation registries instead.
Thinking of asking ChatGPT whether your home blood pressure cuff is accurate? Think again. A new study presented at the American Heart Association's Hypertension Scientific Sessions found that AI search tools are unreliable when it comes to identifying whether home BP monitors have been clinically validated — and the consequences could be serious.
Researchers tested four major AI platforms — Google Gemini, Microsoft Copilot, ChatGPT, and Perplexity AI — on a randomized sample of 324 home BP devices (both validated and non-validated). The results were underwhelming across the board, with most tools performing only marginally better than a coin flip. Notably, AI tools were better at flagging non-validated devices than correctly identifying validated ones.
By the Numbers:
Why it matters: Inaccurate BP readings from unvalidated devices can lead to misdiagnosis or inappropriate treatment decisions. Clinicians and patients should skip the AI shortcut and consult official validation registries like Stride BP, Validate BP, or Hypertension Canada — or simply bring the device to a clinic visit for a side-by-side comparison.
