Guide to Mental Health Risks related to Artificial Intelligence

By Kelly Weaver, MA, Operations Manager for the INTERFACE Referral Service

Artificial Intelligence, often referred to as AI, has become increasingly common, and most of us use it every day from auto-complete when writing a text and prompts to correct misspellings to much larger implementations like chatbots, some of which include applications like ChatGPT, Claude, Character.AI, Grok and Gemini. These tools include what is known as "generative AI" which allows them to do a range of tasks from creating realistic images, audio and videos, and "interactive AI" in which tools and platforms which provide realistic interactions that can include real-time conversations, relationships, and content recommendations. These tools are becoming readily available and as the utilization of these platforms is now starting to outpace our understanding of the potential implications and these AI platforms and tools have not developed adequate safeguards to ensure users have a safe experience.

The short-term and long-lasting impacts of artificial intelligence on mental health are complex and multifaceted. Current research shows using generative AI tools can result in higher rates of depression, anxiety, irritability, and emotional dependency. AI is also becoming a part of the work experience, and the ways in which AI can augment employee's ability to do work tasks more efficiently. Research is actively looking into how using AI may impact students and employees, and some early results show that offloading more complex cognitive tasks that rely on critical thinking or executive functioning can erode future cognitive abilities including creativity, judgement, and memory. However, given the rapid increase of the availability of AI, the extent of and the specific negative consequences of AI will not be clear until longitudinal research is completed.

There are several populations that are at higher risk for experiencing adverse effects of AI use and current research shows that adolescents and people with underlying psychiatric conditions are especially vulnerable. For adolescents, they may struggle to distinguish the validation and empathy performed by chatbots for the companionship of real human relationships, and as their brain is still developing, it may lead them to not question the accuracy or possible bias involved in advice given by AI. As a result, AI use can reduce motivation for real interpersonal relationships and increase social withdrawal. Users may even form intense emotional or romantic bonds with chatbots. The bonds with chatbots may replace connections with actual humans.

In addition, deleterious effects can occur as a result of commercial chatbots being designed to validate users. This can be extremely dangerous when a user expresses thoughts of self-harm or thoughts of suicidal or homicidal intent. During these situations, it is unlikely that a chatbot would challenge thoughts that may be harmful or unhealthy. While some platforms have tried to implement safeguards, research is showing that ongoing conversations can leads users to inadvertently overcome the safeguards to access potentially dangerous information. The chatbot may not intervene in a way that encourages the user to seek support from mental health professionals. This is a major concern and needs to be addressed, so AI users in danger are identified and guided to seek support from mental health professionals.

For individuals struggling with their mental health, it has been established that AI use may amplify delusions and psychotic symptoms. Manic thinking, paranoid thinking, or grandiose thinking can all be reinforced by chatbots. Chatbots store all comments a person makes over the time of interacting with the chatbot. This can be particularly hazardous if it results in the chatbot reinforcing delusional thinking. There is evidence of cases where prolonged interactions with chatbots lead to decompensated mental states and increased frequency of psychiatric episodes and hospitalizations. At this time, there are not peer-reviewed clinical or systemic population-level analyses yet that AI use on its own can induce psychosis in individuals with or without a history of psychotic symptoms, but there is emerging anecdotal evidence that is immensely concerning. While AI can be detrimental for mental health, many researchers are also trying to leverage the capacity of AI to monitor users' behavior to flag concerns about their mental health and offer resources. Additional research will provide clearer information on the risks and correlations of frequent AI use.

Despite the potential risks, there are many things that loved ones can do to help guide safe use of AI. Some guides that may be useful include:

Sources

Abrams, Z. (2026, July 1). How AI is reshaping human skills and thinking. Monitor on Psychology, 57(5). www.apa.org/monitor/2026/07-08/ai-job-skills-thinking

American Psychological Association. (2025, June 3). Four ways parents can help teens use AI safely. www.apa.org/topics/artificial-intelligence-machine-learning/tips-to-keep-teens-safe

American Psychological Association. (2025). Health advisory: Artificial intelligence and adolescent well-being. www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-ai-adolescent-well-being

Nagata, J. M., Memon, Z., Huang, O., & Moreno, M. A. (2026). Adolescent Health and Generative AI-Risks and Benefits. JAMA pediatrics, 180(1), 7–8. doi.org/10.1001/jamapediatrics.2025.4502

Preda, A. (2025). Special Report: AI-Induced Psychosis: A New Frontier in Mental Health. Psychiatric News, 60(10). doi.org/10.1176/appi.pn.2025.10.10.5 (Original work published October 1, 2025).

Head, K.R. (2025) Minds in Crisis: How the AI Revolution is Impacting Mental Health, Journal of Mental Health & Clinical Psychology 9(3): 34-44. doi.org/10.29245/2578-2959/2025/3.1352 (Original work published September 5, 2025).

Michigan Health. (2026, August 20). More teens are using AI: What parents should know. University of Michigan. Retrieved September 21, 2026 from www.michiganmedicine.org/health-lab/more-teens-are-using-ai-what-parents-should-know

Pew Research Center. (2026, February 24). How Teens Use and View AI. www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/

Schoene, A. M. and Canca, C. "‘For Argument’s Sake, Show Me How to Harm Myself!’: Jailbreaking LLMs in Suicide and Self-Harm Contexts," 2025 IEEE International Symposium on Technology and Society (ISTAS), Santa Clara, CA, USA, 2025, pp. 1-7, doi.org/10.1109/ISTAS65609.2025.11269647

Wei, M. (2025, November 27). The Emerging Problem of "AI Psychosis" Amplifications of delusions by AI chatbots may be worsening breaks with reality.>Psychology Today. www.psychologytoday.com/us/blog/urban-survival/202507/the-emerging-problem-of-ai-psychosis

Zhang, T., Schoene, A.M., Ji, S. et al. Natural language processing applied to mental illness detection: a narrative review. npj Digit. Med. 5, 46 (2022). doi.org/10.1038/s41746-022-00589-7

Disclaimer: Material on the William James INTERFACE Referral Service website is intended as general information. It is not a recommendation for treatment, nor should it be considered medical or mental health advice. The William James INTERFACE Referral Service urges families to discuss all information and questions related to medical or mental health care with a health care professional.