I have interacted with a few AI chatbots, and they were quite well implemented. I can't recall what it was, but I think it was KPN or the NS (Dutch ISP or the Dutch railroads).
To start with, it was made clear that it was a chatbot, and you were told to state the question in simple terms.
After a few answers that did not resolve my problem, I was asked to wait and transferred to a human second liner, thus a CS agent with more knowledge and possibilities than a first liner and who was able to resolve my issue.
So, that I am not so sure that using AI, even very simplistic chatbots, will be a bad thing if what it will do is replacing a first line agent who's usual task is to filter out questions and reassign to the appropriate second line or do some simple tasks like resetting a password.
An AI chatbot would actually not be replacing a human agent, but just functioning as an obligatory and fancy interface to the FAQ or knowledge bases that almost everybody has, but users just refuse to use. And its real value would be to avoid that the second and third line support gets overwhelmed by useless queries and can spend more time answering and resolving complex questions.
I have, BTW, quite a few years of experience in the sector.
In a nutshell: Whether an AI chatbot is a good or a bad thing depends, as usual, on the company itself and how they use it.