Healthcare AI Safety Gap · Don't break patients
    Reputable report / news11 Nov 2026·US

    AI chart summaries left out or got wrong diabetes results for many patients at a safety-net system

    Commonwealth Fund: NYC Health + Hospitals found AI-generated chart summaries omitted patients' A1c results nearly 25% of the time and gave wrong information 15% of the time, with omissions three times as common in patients with five or more conditions.

    15%
    of AI chart summaries with wrong A1c information (omitted in nearly 25%)

    What the research found

    Analysts at NYC Health + Hospitals, the largest US municipal safety-net system, evaluated large language models in their electronic health record over the past year. AI-generated chart summaries omitted A1c information for nearly 25% of patients and included wrong information for 15%, and omissions were three times as common for patients with five or more comorbidities as for those with diabetes alone; they did not see significant variation by race or ethnicity. AI-drafted portal replies were written at a 9th-grade reading level against 7th grade for the system's providers.

    Why it matters for providers

    Patients with the most complex conditions were the ones most often missing key results in AI summaries, so providers should test summary tools on their sickest patients before relying on them.

    Original source
    The Commonwealth Fund (NYC Health + Hospitals)
    Read the full report ↗
    A reputable institutional report or news outlet.