People who regularly use artificial intelligence chatbots for emotional support and personal advice are more likely to report symptoms of depression and anxiety, according to a US study published in JAMA Network Open in January that surveyed a large number of people.
The study, by researchers at Mass General Brigham, a Boston-based nonprofit academic health system affiliated with Harvard Medical School, surveyed 20,847 adults across the United States about their use of artificial intelligence and recent experiences with mental health.
Approximately 10.3% of respondents said they use AI tools at least once a day, and 5% said they interact with such systems multiple times a day. Among everyday users, almost half used AI for work and about 11% used AI for school-related tasks. However, the majority (87.1%) said they also use AI for personal reasons, such as seeking advice, recommendations or emotional support.
The average age of participants was 47 years. Researchers found that people who used chatbots daily for personal reasons were more likely to report moderate or higher symptoms of depression, anxiety, and irritability compared to people who did not use AI for other purposes. Symptoms assessed included sleep disturbances, difficulty concentrating, changes in appetite, and thoughts of self-harm.
This association was particularly strong in people aged 45 to 64. In contrast, the use of AI for work or academic purposes showed no association with symptoms of depression.
The researchers observed a phenomenon known as a “dose-response” effect. This means that the more often individuals use AI in their personal interactions, the stronger the association with reported mental health symptoms.
Experts say the study shows a correlation, not cause and effect. People who already experience depression or loneliness are more likely to seek interaction through AI systems, potentially creating what some researchers call a feedback loop.
