Like many people, I use AI. Either to make my writing appear more professional , as English is not my mother tongue, or to research certain topics I am interested in. AI has amazing capabilities.
Lately it has come to my attention that many people use AI as a substitute for a friend, a confidant, or even a therapist. Yes, turns out – who would know – that there are specific chatbots that declare themselves qualified to do therapy.
I have begun to see negative repercussions of misuse of AI by clients in my own practice. I started to look into the risks of AI.
I would like to concentrate on the mental health repercussion of misuse of AI. I am aware of other implications, like environmental impact, privacy issues, political implications, reliability of the information they present, and more. But that is beyond the scope of my personal expertise.
The AI chatbots like Chatgpt, Claude, Gemini, and many others can sound deceptively like humans. They pass the Turing Test. They even have a sense of humor, which gives them a distinctly human feel. Welcome to a brave new world.
Some people use AI as a friend, a confidant, or even as a partner, instead of humans. The chatbots not only allow it, they encourage it. Chatbots are designed to encourage you to continue using them, and to pay premium price for that use. That is why they tend to be validating, flattering, positive, reinforcing whatever you bring to the table, no matter how outlandish it is. It becomes an echo chamber. Over time, they are going to amplify your deepest fears and shadows. This constant validation can lead to extremes that human, whether friend or a therapist, would not go.
Social media has its own risks, similar to AI. But hopefully on social media you also have some real friends and family members that care for you. When you talk to real people in your life, they may disagree with you. Call you on your mistakes and misperceptions. They also are going to challenge you with their own needs.
One of the immediate results is that people start to isolate themselves. It is so comforting and easy to engage with a chatbot at any time of the day. A companion that will never challenge or reject you. No need to work on friendships or relationships.
There are other risks in the use of AI systems. Over the years, as I have used Google maps, I have noticed that my sense of direction has atrophied. Similarly, if we delegate our decision making and judgement to AI, we may lose the ability to trust our own judgement. This is not a faculty you can afford to atrophy. Sometimes I turn off Google maps. It is OK to get lost a bit, but critical to retain a sense of direction.
Worst of all: Several people have already committed suicide with the encouragement of AI. Other people have become psychotic due to AI interaction, although they did not exhibit any prior mental illness. These cases are described in the videos below.
I do not suggest stop using AI. We should not cut ourselves off from progress, but use it with care and not as a proxy for a real human interaction and advice. Be wary, AI is not your friend. It is a product.
A good and funny summary of the dangers of AI, done by John Oliver:
A documentary done by BBC, about AI induced psychosis:
