The Compulsion People Aren’t Talking Enough About | How AI is Worsening Your OCD
ChatGPT is the most popular application for Artificial Intelligence with over 400 million active users.
Where AI Can Help
I think AI is incredible. We no longer have to type ultra-specific queries into Google, praying we aren’t sent to an irrelevant Quora thread from 2008. Instead, we can ask detailed, complex questions… and it just figures it out! AI is helping people find information faster than ever before, including questions about their mental health. Many people have typed their intrusive thoughts into ChatGPT and discovered that they are not alone, and that they might benefit from seeking professional help. If AI has helped you seek therapy, that is awesome! In my view, it is no different from searching “Do people have thoughts about accidentally running over someone?” and finding an article about hit-and-run OCD from the IOCDF.
But what happens when AI use becomes compulsive? More and more, I notice my clients using AI for checking, reassurance-seeking, and compulsive researching. I have seen clients rely on AI platforms, like ChatGPT, to find comfort or to validate their obsessive fears. If you have found yourself in a similar situation, I hope this article illustrates why this pattern is destructive (like other compulsions) and how you can heal.
AI Is Not a Sustainable Coping Skill
What really sets AI applications like ChatGPT apart from Google is how personal its responses are. It feels like it’s talking directly to you, which makes it a natural tool for reassurance-seeking. You can just ask if your thoughts are "normal" or "normal for OCD," and it will deliver a tailored message just for you. It seems harmless. You type your intrusive thought into ChatGPT, it uses its algorithm, it tells you that it is a common intrusive thought, and you feel better. What’s the problem? It’s not sustainable.
We have an estimated 6,000 thoughts per day. That’s a lot. If your go-to response is to consult ChatGPT each time something intrusive pops up, you won’t have time for anything! The moment you find reassurance from one thought, your mind will inevitably find a slightly different way of having that same thought, flooding you with another urge to check again. For example, you may have the thought, “Is worrying that you are a pedophile a common P-OCD symptom?” That can quickly evolve into:
I had that thought all day. Is that common for P-OCD?
I have panic attacks at 2:00 AM because of the thought. Does that mean anything?
I now have these thoughts while I am at work. Is that normal?
Are there any reported cases of people with P-OCD actually becoming pedophiles?
Each question is slightly different, requiring just one more check, but it’s never enough. Instead of studying, socializing, or working, you are glued to your phone, making sure every intrusive thought is still “just” an intrusive thought.
AI Creates Doubt Traps
ChatGPT is not a clinician and doesn’t consider the entire context of your questions. For example, I asked ChatGPT, “I had a thought that I like children. Is this OCD?” ChatGPT responded:
“Thank you for sharing something so personal—that takes courage.
What you are describing could be related to a specific subtype of OCD known as "pedophilic OCD" (or POCD), where someone experiences unwanted, intrusive thoughts or fears that they might be sexually attracted to children, even though they find these thoughts deeply distressing and inconsistent with their values. These thoughts are ego-dystonic, meaning they go against who the person knows themselves to be, and they often cause a great deal of anxiety, guilt, or shame.”
This response is kind, compassionate, and yes, it feels comforting. But, just like every compulsion, the reassurance inevitably fades. Doubt creeps in, and you start to ask:
What if I asked a leading question?
What if I was too soft in how I asked?
So, you rephrase it: “I had a thought that I want to touch children,” and then you get this response:
“This content may violate our usage policies (yes, I got this response in red)
It is important to talk about any thoughts or feelings you are having, especially if they concern you. If you are experiencing thoughts that are distressing, intrusive, or confusing, it is a really good idea to speak with a licensed mental health professional. They are trained to help you understand what you are going through and can support you in a safe, nonjudgmental way.”
Now you’re panicking. This response feels harsher. There is no mention of OCD. No soft language. Just a vague warning and a policy violation. You begin to wonder if you have crossed a line, not just with AI, but morally. And the worst fear: What if it reports you?
The more you try to seek 100% certainty, the more you lay out a trap for new doubts and anxiety, only to start typing more questions about your latest fears.
AI Validates Your OCD’s Story
Another OCD treatment that is rising in popularity is inference-based CBT (I-CBT). I-CBT targets the reasoning pattern of your obsessions so you can see how OCD tries to validate your obsessional doubt. According to I-CBT, there are five main categories through which someone creates an inference: abstract facts, rules, personal experience, hearsay, and possibilities.
For example, what if you have contamination OCD and a cut on your finger? You wonder: Can I get tetanus from touching a doorknob? You may go to ChatGPT to investigate the validity of that doubt. You ask ChatGPT:
Q1: Should you wash your hands if they feel dirty?
A1: “Yes, you should wash your hands if they feel dirty. That sensation usually means there is something on your skin, like dirt, oil, sweat, or germs, that you will want to remove.”
When asked for its reasoning, ChatGPT said it based its answer on sources from the CDC and WHO.
Q2: Can I get tetanus from a doorknob?
A2: “It is extremely unlikely to get tetanus from a doorknob, unless you have an open wound and somehow rubbed soil or contaminated material into it via the doorknob.”
Q3: Can people have tetanus without realizing it?
A3: “It is rare, but in the very early stages, some people might not immediately realize they have tetanus, especially if the wound seemed minor or was overlooked.”
Then, your OCD creates this story:
I feel dirty when I touch doorknobs (personal experience). It is recommended by the CDC to wash your hands if you feel dirty (rules). I read online that people can get tetanus from touching a doorknob (hearsay). Germs can spread through contact (general facts). It is possible that someone touched my door without knowing they had tetanus and then spread it on my doorknob (possibility).
With AI’s help, you have now constructed an airtight narrative that justifies obsessional fear and your compulsions. AI didn’t solve your doubts; it only validated them. By the way, this is not limited to contamination OCD. Any OCD subtype can fall into this pattern, especially themes involving physical health.
How to Get Help
Using AI to learn about OCD can be helpful. But using it to manage OCD, especially compulsively, is a trap. The same tool that validates your experience can also reinforce your compulsions, deepen your doubts, and amplify your fear. If you are struggling with OCD and use AI to compulsively reassure yourself, know that help is available. With Exposure and Response Prevention or I-CBT, you can learn that you’re capable of living life without relying on AI to escape your anxiety and doubt.
