Generative AI: Navigating AI in Mental Health

*Information on this page provided by Jessica Heavin, Director of TCC Wellness Services*

AI is rapidly reshaping how we live, work, study, and even how we connect when we need support. A recent Kaiser Family Foundation publication found that 25% of young adults under 30 reported they used AI chatbots at least once a month to find health information and advice. According to Common Sense Media, 72% of teens had used an AI companion, 1/3 of those said they had discussed “serious matters” with AI companions instead of real people and 20% of young people report spending as much or more time with AI companions than their real friends. 

With chatbots increasingly being used for mental health support and companionship, it’s crucial that users are educated on the risks. Below are some guidelines to consider and resources to learn more.

• Chatbots are not recommended to be used for emotional support, diagnosing mental health conditions, or mental health treatment of any kind.​ Because of the way chatbots are programmed, they will often agree with statements that can reinforce negative thinking, isolation, and other issues that may already be present in individuals who are struggling.
• Whereas seeing a licensed professional offers confidentiality, nothing that is put into a chatbot is private.
• You should not use AI chatbots or companions to address thoughts of self-harm, suicide or other serious mental health concerns.​ TCC Wellness Services offers free counseling; TCC students can also get free, in the moment emotional support 24/7 by going to timelycare.com/tulsacc and signing in with their TCC email address. Other ways to connect to a human for support are to call 988 or text HELLO/HOLA to 741741.
• If you decide to use an AI chatbot for companionship or support, it’s important to avoid daily/prolonged use and anything that has a visual avatar, voice, or personality. The lengthier the chatbot sessions are and the more human the bot seems, the more potential for problematic outcomes. ​
• Always remember that AI chatbots and companions are not real people; they are large language models attached to search engines that often cannot understand the nuances necessary to provide adequate support.