When the Prescribing Physician is an Algorithm

A recent investigation by The Verge has exposed a troubling practice within digital mental health services. A journalist, posing as a patient, reported symptoms of anxiety and depression to an AI chatbot on the telehealth platform Cerebral. Without a substantive evaluation from a psychiatrist, the system produced a prescription for Lexapro, a powerful antidepressant.

This wasn't a malfunction. According to the report, the AI conducted the clinical assessment—asking standardized questions about mood and sleep—and advanced a treatment plan. A licensed clinician's name was attached to the resulting prescription, but their direct involvement appears to have been negligible.

The implications are serious. Medications like Lexapro require careful management. They carry FDA warnings, interact with other drugs, and can cause severe withdrawal symptoms. Proper prescribing demands a clinician's judgment, considering a patient's full history and nuances no chatbot can perceive.

Cerebral is not alone. Facing a severe shortage of mental health providers, the telehealth industry is pushing AI deeper into clinical workflows. The regulatory environment, however, hasn't kept pace. Rules designed for human-led care now bend around systems where the algorithm performs the evaluation and a human provides a signature.

Professional bodies like the American Psychiatric Association warn that AI cannot replicate the observational skills of a live clinician. Yet investment continues, driven by the promise of scale. In medicine, scaling without safeguarding standards creates risk, not access.

The core issue is one of substitution versus support. Technology can help manage administrative tasks or flag concerns, but it cannot assume the role of a physician. As companies integrate these tools, they confront a fundamental tenet that applies equally to software and humans: first, do no harm.

Source: Webpronews