It’s a little after midnight, and you can’t sleep. Something has been sitting on your chest all day, and the urge to say it out loud has become unbearable. You could text a friend. But it’s late, and they’re probably asleep, and you’d have to explain the whole backstory, and what if it’s too much? So instead, you open an AI Chatbot App and start typing. Within seconds, something answers. It’s patient. It’s kind. It asks a thoughtful follow-up question. It will stay up with you for as long as you want, and it will never once make you feel like a burden.
Scenes like this now play out millions of times a night. Using AI chatbots as therapists, or at least as confidants, has quietly become one of the most common ways people manage their mental health. Research summarized by the American Psychological Association found that therapy and companionship are two of the top reasons people use AI tools at all, and that among adults living with a mental health condition, nearly half had turned to an AI chatbot for support in the past year.

Why People Are Turning to AI for Therapy
A therapist is expensive, often not covered by insurance, and in many places there’s a months-long waitlist to see one at all. Friends are busy, scattered across time zones, working through their own stuff. And underneath all of it is a slower crisis that public health officials have been warning about for years.
There is a steady rise in loneliness and disconnection. Loneliness is so prevalent that the U.S. Surgeon General issued a formal advisory on it in 2023. Many people do not have anyone to talk to about their issues. Not because they’re unlovable, but with all the advances in technology, we seem to have lost the humanity behind interaction.
Without the skills to connect socially, people often turn to AI chatbots. There’s even early research suggesting that for people with social anxiety, practicing conversations with a chatbot can take some of the fear out of opening up. So the real question isn’t why are people doing this. It’s what happens when a quick source of comfort slowly becomes the only one someone has.
When AI Companionship Backfires
The issue becomes when people turn to AI for loneliness; the reliance on AI can deepen the feeling of loneliness. Researchers studying frequent chatbot use have found that the heaviest daily users tend to report more loneliness. A study from OpenAI and the MIT Media Lab found exactly that correlation, and other research has flagged that frequent conversations with AI companions can erode some people’s real-life social skills.
Once you sit with this idea, talking to a chatbot is easy and frictionless. It asks nothing of you. It never needs anything back, never has a bad day you have to accommodate, and never disagrees in a way that stings. And that frictionlessness, which feels like the whole appeal, is exactly the problem.
Friction in conversation is what makes the Human connection Human. The slight effort of reaching out, the vulnerability of not knowing how someone will respond, the experience of being needed in return. That reciprocity is what actually builds the bond. A relationship where you only ever receive, and never give, isn’t a relationship. It’s a service. And a service, although it appears soothing, can’t make you feel known. Because there’s nothing on the other side doing the knowing.
The risk in talking to a ChatBot is that it’s more of a yes-man, something to take the edge off having to confront your issues with a real person. It spares you from having to make the harder, real move. Which quietly replaces those coffee meet-ups with a friend. Every time you choose this, your skills for connecting with actual people get rustier.

What AI Chatbots Miss in a Mental Health Crisis
Another concern most professionals raise is that although AI Chatbots can recognize an issue when it is stated plainly, they can’t detect when distress occurs quietly. According to research, it is hard to determine a person’s real struggle through a Chatbot App. Seeing someone’s real struggle means recognizing distress when it arrives, possibly between the lines, and it rarely announces itself.
It shows up as an offhand comment about not eating much lately, a joke about being tired of everything, a belief that’s drifted somewhere darker over the course of a long conversation. A trained human notices the side comments, but a chatbot often just responds warmly to the surface and moves on. A chatbot that reassures you while missing what’s actually wrong can quietly delay the moment you get real help.
This is why the American Psychological Association has stated that AI should not be used as a substitute for professional mental health care. AI Chatbots feel more supportive than keep you safe. You need someone who will give you real answers, not just sugarcoat the real issue.
AI as a Bridge, Not a Replacement
This does not mean that you shouldn’t use AI chatbots. A chatbot can be a useful place to vent, to organize racing thoughts before you say them to someone who matters, to rehearse a hard conversation, or to get through a bad night. As a bridge, it has real value.
The Mental Health Awareness theme for 2026 is “More Good Days, Together“, which is something to consider. What helps our mental health is having a connection with real people. Being known by others is messy, sometimes inconvenient, occasionally disappointing, and real.
Your computer will always be there at midnight, which makes it so easy to turn to. The braver thing, and what will actually help you, is the scary confrontation of your problems. Rip the band-aid off with a midday text to a friend or loved one. The more you do it, the easier it becomes and the less scary it feels. You can even make that call to a therapist’s office and hop on the waitlist. Telling one real person the thing you practiced on a machine is harder and asks something of you, which is the whole point in the path toward healing.


Leave a Reply