Study Links Use of AI for Health Questions to Higher Anxiety and Depression in Young Adults

The study found that individuals using AI tools like ChatGPT or Claude for health-related questions had a 52% increased likelihood of screening positive for anxiety and a 46% increased likelihood of screening positive for depression compared to non-users. While a correlation between AI use and mental health issues was established, the study does not conclude that AI usage causes these conditions, nor that individuals with anxiety or depression are more inclined to seek health information from AI.

Zhai noted the significance of these findings, especially as discussions about AI in education often focus on its use for academic tasks. His research aims to understand how technology can identify and address mental health risks. He explained that many young people may turn to generative AI when feeling anxious or unsure about health symptoms. This raises critical questions about whether repeated engagement with AI for health queries could lead to increased health anxiety or distress.

In the study, approximately 52% of generative AI users screened positively for anxiety, in contrast to 43% of non-users. For depression, the figures were around 47% and 38%, respectively. These associations held even after accounting for various demographic factors such as age, race, sex, and socioeconomic status.

Zhai highlighted a potential pattern where uncertainty about health symptoms prompts individuals to seek reassurance from AI, leading to additional questions and possibly creating a cycle of rumination. He emphasized that generative AI's conversational nature distinguishes it from traditional internet searches, which typically only provide links to information rather than engaging directly.

He cautioned that the confident responses generated by AI could mislead users, particularly when information is inaccurate or misleading. Zhai emphasized the need for greater awareness among young people regarding how they use AI for health inquiries and the emotional impact of these interactions.

He recommended mental health professionals discuss AI use with clients to assess the nature of the information being sought and to assist in evaluating its credibility. Additionally, universities could integrate guidance on effective AI use into broader AI literacy programs, helping students discern reliable health information.

Zhai concluded that the core issue is not generative AI itself but rather how individuals choose to use it, reinforcing the importance of understanding the implications of AI interactions on mental health.