Researchers at SINTEF and the University of Oslo have published a study in DIGITAL HEALTH in which young people consistently rated ChatGPT's responses to real mental health questions as more useful, more relevant, and more empathetic than answers from health care professionals.
The study compared answers to genuine queries that young people had submitted to a Norwegian youth mental health charity, with 123 young people and 31 professionals reviewing both sets of responses. Both groups gave ChatGPT the highest grades, with young people more enthusiastic than the professionals. The researchers' explanation is sobering. Health professionals are constrained by guidelines that limit them from giving direct advice or diagnoses outside a clinical relationship. ChatGPT operates with no such constraints, which makes its answers feel more immediately useful, even when a clinician would consider the same advice incomplete or unsafe.
