If there's an "AI effect" surely its inverse also exists; it seems to me that every time there is a [relatively] major advancement, a similar chorus of supporters pop up to tell us how we're so close to solving every AI problem imaginable. In recent memory there's Google's NMT, which had people effectively saying it's "solved" translation, but when I got to actually feed something into it, as a competent speaker of Japanese and former translator, I can only say that I was wholly unimpressed, relative to the hype at least.

Discounting AI advances is nothing new, but upselling them is at least as old, as well.