Idea for Protecting the Company and the Users

I’d love to see ChatGPT add an optional daily AI relationship check-in.

The idea: when a user signs in, the system could briefly review patterns from the previous day or two and surface a short, private reflection about how they’ve been using AI.

Not a diagnosis. Not a scary “health score.” Not a warning every time someone uses ChatGPT a lot.

Instead, it could notice patterns such as:

  • repeated reassurance-seeking
  • increasing emotional dependence on the chatbot
  • repeatedly asking AI to make important personal decisions
  • signs that AI is beginning to replace human interaction rather than supplement it
  • escalating reality-testing concerns
  • unusually intense or rapidly increasing use

It could also acknowledge healthy patterns: learning, creating, planning, exploring ideas, improving work, or using AI to consider multiple perspectives.

What makes a chatbot uniquely suited to this is that it could phrase the feedback in a way the individual user is likely to hear. The underlying safety criteria could remain consistent, while the tone and explanation adapt to the person.

For example:

“You’ve returned to the same worry several times over the last two days. It may be worth considering whether continuing to ask for reassurance is helping you, or keeping the worry active.”

The goal would not be to diagnose users or restrict access. It would be to make gradual changes in the human-AI relationship visible to the user before potentially unhealthy patterns become severe.

I think this could benefit both users and AI companies. It would create a form of proactive safety rather than relying primarily on intervention once a conversation has already reached a crisis point. It would also give companies a concrete way to demonstrate that they are actively trying to identify and reduce unhealthy dependency on their systems. People can struggle to monitor their own dependence, so why not help them?

AI has enormous potential to improve people’s lives. As these systems become more conversational and personally useful, I think helping people maintain a healthy relationship with them should be an integral part of the product design from the beginning.