A study by Brightside Health shows that large language models (LLMs) can accurately predict mental health crises like suicidal ideation. The study compared the accuracy of LLMs to clinicians and found that LLMs predicted with 61.5% accuracy while clinicians predicted with 55.2% to 67% accuracy. LLMs were also significantly faster at evaluating patient data, taking less than 10 minutes compared to over 3 hours for clinicians. With a shortage of mental health providers, AI has the potential to support clinicians in identifying at-risk patients efficiently. However, Dr. Winsberg emphasizes the importance of using AI as a tool to assist clinicians, not replace them.
Source link
Study: LLMs Identify Mental Health Crises with Accuracy Comparable to Clinicians
