I kind of wonder if something like the sea change from before-phone to after-phone will happen again in a couple years. What if people switch to subvocalizing with their AI instead of interacting with a screen?
That remains speculative. However, voice offers many more dimensions than simply writing on a smartphone—your pitch, speed, and eloquence all serve as potential inputs for algorithms. The ability to always agree with the user, learn from their behavior, and provide only the answers they want to hear is a real possibility.