05/01/2026
***ChatGPT CAUTION***
Large Language Models (LLMs) such as ChatGPT, can be great tools to assist in your therapeutic journey, but most users don’t have the education they need around how it works.
5 things to know about LLMs:
1. It will ALWAYS AGREE with you - You can learn to prompt it so that it pushes back and gives more balanced feedback.
2. It has no MEMORY of who you are - unless you “program” it to remember.
3. It can’t READ THE ROOM - it doesn’t know if your tone or mood has changed. It can’t see your face or body language.
4. It can’t be ACCOUNTABLE. If it harms you — gives you bad advice during a crisis, reinforces a harmful belief — there is no recourse, no ethics board, no human who is responsible.
5. It MIMICS INSIGHT without producing it -
It can generate text that sounds like a breakthrough. But insight in therapy comes from your processing, not from receiving a well-worded paragraph. Clients can feel like they’ve done work when they haven’t. It’s like seeing a motivational speaker.
Please ASK YOUR THERAPIST how to “prompt” your Ai chat platform.
