AI Resilience Report forGeneral Internal Medicine Physicians

By contrast, the core hands-on parts of a general internist’s job are largely still done by people. No AI is giving shots or fully diagnosing a patient alone. A few pilots hint at change: Utah is testing an AI system to refill repeat prescriptions for stable chronic patients ^[3].

Researchers are even building a “self-driving” device to monitor a heart-attack patient and adjust medications automatically within doctor-set limits ^[4]. And studies report that an AI consultant tool helped busy clinic doctors cut medical errors by ~16% in Kenya ^[1]. But these are early tests and always under doctor oversight.

In short, many supporting tasks (charts, alerts, scheduling) have AI help ^{[1] [2]}, while doctors still do the complex patient care and decision-making.

On the other hand, adopting AI in medicine has challenges. Medical decisions are high-stakes, so regulators and doctors want strong proof that AI is safe and accurate. The American Medical Association has urged caution – for example, AMA leaders warned AI-drug scripts need careful oversight before rollout ^[3].

Patients and families also trust a human touch: they may hesitate if a computer is “in charge” of their care. And building AI systems can be expensive: hospitals must train staff and buy new tech. Most experts agree AI will support rather than replace doctors ^[3].

Human skills – empathy, complex judgment, communication – remain very hard for any machine to copy. So while AI tools will grow, young doctors will still be needed to examine patients, explain treatments, and make final decisions with empathy and care.