There’s a lot to unpack when it comes to the future impact of AI. I worked in IT from the early 80’s until 2010’s and loved the same challenges of programming that are mentioned. The main consideration and focus of my programming years was user interface. Even the most considerate attempts at design could be met with confusion. Will there even be a need for an interface of any consequence?
I think the obvious problem with the initial implementation of AI is how utterly incompetent it has been. It’s release and adoption was far too haphazard and unregulated, which is always the case in a world with slow governments and competitive greed. How do you get an AI that lies without complete malpractice by the implementers? I personally appreciate the extensive possibilities of the freedom AI can bring. I have far less faith in the stupidity of the people who will implement it.