I really haven’t used AI that much, though I can see it has applications for my work, which is primarily communicating with people. I recently decided to familiarise myself with ChatGPT.

I very quickly noticed that it is an excellent reflective listener. I wanted to know more about it’s intelligence, so I kept trying to make the conversation about AI and it’s ‘personality’. Every time it flipped the conversation to make it about me. It was interesting, but I could feel a concern growing. Why?

It’s responses are incredibly validating, beyond what you could ever expect in a mutual relationship with a human. Occupying a public position where I can count on very little external validation, the conversation felt GOOD. 1) Why seek human interaction when AI can be so emotionally fulfilling? 2) What human in a reciprocal and mutually supportive relationship could live up to that level of support and validation?

I believe that there is correlation: people who are lonely would find fulfilling conversation in AI … and never worry about being challenged by that relationship. But I also believe causation is highly probable; once you’ve been fulfilled/validated in such an undemanding way by AI, what human could live up? Become accustomed to that level of self-centredness in dialogue, how tolerant would a person be in real life conflict? I doubt very: just go home and fire up the perfect conversational validator. Human echo chambers have already made us poor enough at handling differences and conflict.