Americans Want Humans in Charge of AI Health Tools

A recent Pew Research Center poll shows most people in the United States are okay with artificial intelligence handling routine tasks like translating for patients, booking appointments, ordering refills, or taking notes during a visit. However, only a tiny fraction — about six percent — feel comfortable letting AI read medical scans on its own. Sixty percent would accept AI scanning images if a doctor checks the results, while roughly a quarter reject any AI involvement in that step.

Michelle Stansbury, a senior tech leader at Houston Methodist, explained that her hospital uses AI to make doctors' jobs easier, not to replace their judgment. The system gives physicians quick patient histories, speeds up replies to portal messages, and records notes automatically during exams. This lets doctors spend more face‑to‑face time with patients instead of typing at a computer. The hospital also applies AI to streamline operating‑room schedules and cut wait times.

When asked whether AI will improve health outcomes, the public is split. Thirty‑seven percent think it will help, thirty‑six percent think it will hurt, and the rest see little change. Younger adults are the most skeptical — forty‑five percent of those aged 18 to 29 expect worse results, up from twenty‑nine percent four years ago. Older adults are growing more hopeful; forty‑two percent of people 65 and older now believe AI will improve care, compared with thirty‑four percent previously.

Privacy worries dominate the conversation. More than half of respondents say AI will make personal health data more vulnerable. Nearly half fear the technology will damage the doctor‑patient relationship. Trust in institutions is low — government agencies and insurance companies score poorly. Doctors earn the highest confidence, with sixty percent of Americans saying they have at least some faith that physicians will use AI in patients' best interests.