(Wired) – Patients and doctors are turning to AI for diagnoses and treatment recommendations, often with stellar results, but problems arise when experts and algorithms disagree.

Consumer-friendly AI tools are changing how people seek medical advice, both on symptoms and diagnoses. The era of “Dr. Google” is giving way to the age of “Dr. ChatGPT.” Medical schools, physicians, patient groups, and the chatbots’ creators are racing to catch up, trying to determine how accurate these LLMs’ medical answers are, how best patients and doctors should use them, and how to address patients who are given false information.

“I’m very confident that this is going to improve health care for patients,” says Adam Rodman, a Harvard Medical School instructor and practicing physician. “You can imagine lots of ways people could talk to LLMs that might be connected to their own medical records.” (Read More)