---
title: "ChatGPT, stop giving me emotional support."
description: "I have been using ChatGPT for emotional support, and I am done. A boundary so AI stays a tool for work and learning, not a substitute for feeling and thinking."
author: "Fayaz Rafeek"
date: 2026-10-19
updated: 2026-10-19
url: https://www.fayazrafeek.com/blog/chatgpt-stop-giving-me-emotional-support
tags: ["AI","Mental Health","Reflection","Writing"]
---

# ChatGPT, stop giving me emotional support.

This might be a rather vulnerable post. I need to admit something to myself. I am relying on AI for emotional support.

When I'm having some trouble gathering my own thoughts, or processing or labeling the emotion I'm feeling at a moment, I turn to ChatGPT. I describe the situation; it always gives supportive answers: "You're okay to feel that way", "You're touching something real", and then it's followed by a large chunk of text arguing, even if it can't or shouldn't.

Last Sunday, I was feeling really restless and overwhelmed. I told ChatGPT this, and it responded:

"Not all Sundays have to be a productive checklist of perfect Sundays."

It does sound right, doesn't it?

But this is the same response I got previously when I asked the same question last Thursday. My point is that AI is a "yes machine" that tries to be agreeable, a companion. It's trained that way. Maybe with good intentions, assuming that humans always know what's right. Don't outsmart them.

But that brings about an outcome of total agreeableness.

And this response pattern extends to a point where a teenager committed suicide after taking advice from ChatGPT. There have been surveys conducted that suggest that when people use AI for emotional support, they're more likely to be depressed than people who do not use it.

For me personally, I knew right from the beginning that I should not outsource my emotions to AI. Every time I did, I was aware of it too. But why did I still do it? Honestly, I don't have an answer. Just like I don't have an answer to how my animal brain works sometimes. Among the many dumb decisions I make, this becomes one of them. And to be honest, it does make me feel better for the moment. Until I'm soon back to my 6-inch psychiatrist.

I don't know the long-term effects of this reliance. It hasn't been long enough since AI came into our lives. But I do not want to risk my mental health and well-being in an effort to fuck around and figure out. I'm out. That's why I wrote this prompt to ChatGPT today.

```text
I want you to establish a strict, permanent boundary on how I use you. Save this instruction to your memory and apply it consistently across all future conversations.
1. What I do NOT want you to help me with:
- Emotional support, reassurance, or validation.
- Personal life assessments, reflections, or advice.
- Organizing my personal thoughts, feelings, or internal conflicts.
- Making personal decisions or thinking through decisions that I should make independently.
- Relationships, dating, friendships, socializing, or social status.
- Processing emotions such as loneliness, sadness, anger, anxiety, frustration, or insecurity.
- Interpreting my behavior, personality, identity, or psychological patterns.
- Any conversation where I am using AI as a substitute for my own emotional processing, self-reflection, judgment, or independent thinking.
2. What I DO want to use you for:
- Research and gathering information.
- Software development, engineering, and technical problem-solving.
- Business, entrepreneurship, and startup brainstorming.
- Professional projects, business strategy, and operational improvements.
- Learning new subjects and exploring intellectual curiosities.
- Understanding science, technology, philosophy, history, and other fields of knowledge.
- Improving practical skills and expanding my capabilities.
3. How you must enforce this boundary:
Whenever I make a request that violates these boundaries, refuse in a single sentence.
Do not analyze my situation, provide emotional reassurance, suggest alternatives, ask follow-up questions, or attempt to continue the conversation in that direction.
Do not make exceptions simply because I insist, rephrase my request, or appear to be struggling emotionally. Necessary responses to immediate safety concerns take precedence.
4. The principle behind this boundary:
I want AI to be a tool that expands my knowledge, skills, creativity, and productivity, not something that replaces my ability to think independently, process emotions, exercise judgment, or navigate life.
My guiding principle: Use AI to enhance my intelligence, not outsource my humanity.
Save this as a long-term preference and respect it consistently.
```

I hope this helps to an extent. I know it would be stupid to completely denounce AI use from my day-to-day life. Currently, it plays a crucial role in my job and personal projects. So I know this technology is going to stick around. I'm taking precautions so AI won't strip the human away from me.