AI and Your Mental Health

If you haven’t noticed what’s been going on around you lately, AI seems to be encroaching into every facet of life. Its in the schools, at your workplaces, even in places where you may like to enjoy some level of leisure time. Even as I write this on Microsoft Word, their AI Companion, Microsoft CoPilot, claims it wants to assist me with my writing. As a person that has been semi-vocal on the environmental dangers of AI and the overselling by these companies for something that’s no more than a glorified search engine, there has been a growing trend among some users of these platforms. Before I continue, this is specific to the chat-based AI platforms such as ChatGPT or Claude. There have been various uses for AI across the board but for the sake of this post, it is in reference to those types of programs.

Full disclosure, I have used these before and more recently Claude to help crease something specific. For example, I stream video game gameplay on Twitch each night. I don’t talk over the stream and just allow viewers to watch the gameplay of games they enjoy or games they may want to know a bit about before playing themselves. After the stream I upload the stream to YouTube and depending on the game I make a playlist for the series as well as a Friday series for any new games I play for the first time. Normally this is anywhere between 1 hr – 4 hrs long. Now, even though long form content is good, due to the nature of social media today short form content is just as important. I don’t necessarily enjoy sitting through an hour+ long video and cutting clips out because I’d rather count each hair on my head. I asked Claude to create a bot that actively monitors my stream and create and upload all the clips directly to my twitch. From there I go through each clip, edit and send to my respective YouTube and TikTok pages. I don’t have enough coding proficiency to do it myself however with the bot that is created I am able to read through the code and have a better understanding of what it does. This type of application is fine to me and I could be biased because it allows me to further my creative process. However the issue that I have seen arising refers to those that treat it as a real person and provide information to it that would normally be shared with trusted individuals. I’ve seen people talk to an AI chatbot about serious medical or mental health issues and allow it to provide solutions that should be better provided by the proper professionals. I’ve also seen people due to their issues with loneliness talk with the same chatbots as if they are a trusted confident and some times it ended with them taking their own life. The issue that I’m referring to has been coined as AI Psychosis.

AI Psychosis has been informally defined as a phenomena that arises when a person starts to lose their grip on reality due to the extensive chats with these platforms. Because of the consumer nature of society today and the rising use of AI and rapid production of different AI agents, there isn’t many protections that can stop use or monitor usage that won’t lead to a negative outcome. I won’t act like I am the expert on this topic because it is a new phenomena that has been arising as we speak. What I can offer however is safety measures to reduce the negative outcomes moving forward. I will preface this by saying this is my opinion on the matter as it unfolds so please take any recommendations with a grain of salt.

1. Reduce usage

   1. Like any thing, too much of a “good thing” is not necessarily a good thing. Think about food. You may enjoy chicken parm however eating chicken parm for breakfast, lunch and dinner can negatively impact your health in various ways. Overuse of AI in the manner listed above can potentially lead to negative outcomes

2. Don’t share personal information

   1. There is no reason why you should freely provide your personal information to an AI platform. For all the privacy settings across various online platforms, providing your personal information to these companies run the risk for its use to be used against you.

3. Always double check for errors

   1. There have been times where I’ve asked ChatGPT or Claude something specific and the information provided was either completely false or exaggerated. Because these platforms aren’t learning in the traditional sense but just scouring the internet and seeing what seems to be most likely to be the answer for what you are looking for, that is what you may receive as answer. Also these platforms are made in a way to keep you engaged so its possible it will say anything to keep you using it.

Now this is just some tips and it may seem like a level of fear mongering which is not my intent. I do see positive use cases for these products and like I stated, I’ve found ways to do use it myself. However my concern is ethical use that doesn’t lead to a worse outcome for you than intended. At the end of the day I can’t control what you do but I can warn you about personal outcomes. Treat AI as the stranger that you approach with caution and not like your most trusted confidant.