AI Chatbot Shows Potential in Treating Depression, Study Suggests

Published about 1 year ago
  Pixabay

Artificial Intelligence (AI) chatbot ChatGPT, could be more effective in adhering to depression treatment guidelines compared to human doctors, according to a recent study. The study indicates that the AI technology might enhance decision-making in primary care, as it can adhere to accepted treatment standards devoid of biases related to gender or social class that sometimes occur between humans. However, more research is needed to examine any potential risks or ethical concerns that might arise from its practical use.

Researchers from Israel tested two versions of ChatGPT, 3.5 and 4, using brief descriptions of hypothetical patient scenarios demonstrating symptoms of depression during initial consultations. The scenarios consisted of eight distinct characters differing by gender, socioeconomic status, and depression severity. Symptoms included sadness, sleep issues, and loss of appetite in the three weeks leading up to the appointment, as well as a mild to moderate depression diagnosis.

The researchers used this hypothetical patient data with ChatGPT 10 times and compared its responses to those of 1,249 French primary care doctors, 73% of whom were women.

ChatGPT’s Recommendations vs Doctors’

In cases of mild depression, ChatGPT-3.5 advised psychotherapy in 95% of instances, and ChatGPT-4 recommended the same in 97.5% of cases. In contrast, primary care doctors suggested psychotherapy in only 4.3% of instances, choosing drugs 48% of the time, or a combination of psychotherapy and prescribed drugs 32.5% of the time.

In severe depression cases, 44.5% of doctors recommended a mixture of psychotherapy and prescribed drugs, while ChatGPT recommended this method in 72% and 100% of cases respectively. When it came to the type of medication advised, ChatGPT leaned towards the exclusive use of antidepressants in 74% and 68% of cases, while human doctors favored a mix of antidepressants and anxiolytics/hypnotics in 67.4% of cases.

Findings and Future Research

The study’s findings, published in the journal Family Medicine and Community Health, show that ChatGPT aligns considerably with accepted guidelines for managing mild and severe depression, without displaying the gender or socioeconomic biases observed among primary care physicians.

The researchers pointed out that ChatGPT-4 demonstrated superior precision in adjusting treatment to comply with clinical guidelines, suggesting it has the potential to enhance decision-making in primary healthcare. Nevertheless, despite the potential benefits of utilizing AI chatbots like ChatGPT, more research is required to refine AI recommendations for severe cases and to examine potential risks and ethical issues.

Related news