This "AI" is a canard. Just like an old TI calculator, one doesn't simply mash a bunch of microelectronics together and produce "intelligence". "AI" doesn't invent anything, nor will it have an original idea. Much like the TI calculator, it simply calculates (or in the case of AI, compiles) information far more quickly than any human may do. It doesn't supersede humanity - it simply speeds it up.

Calculators must be progammed, ultimately by humans at the source, to produce mathematical results. Calculators produce results that are logical given the information programmed into it. Calculators wouldn't devise Riemannian geometry on its own - but once programmed, a calculator may produce results that are congruent to those expected by those humans who also produce results using the same branch of geometry.

"AI" is much the same. "AI" simply regurgitates the information provided to it, in an elegant format. Ultimately, "AI" is also programmed by a human (or group of humans) and will advance the beliefs of those humans because it simply follows the programming. "AI" doesn't produce new ideas. "AI" may display conundrums (conundra?) if allowed to do so, but "AI" will be useless to solve those in any way other than the ways that humans have programmed it to solve.

So, why "AI"? It's nothing new - it's just faster. But - in the face of widespread distrust of government and it's premier progranda tool, the media, the rulers need to find a new "Wizard of Oz" that will be "trusted" to provide perfect, "non-biased" solutions and remove all dissent. "Well, the AI said to do this, and the AI is so much smarter than every human being, so AI is right and you are wrong". It is literally a canard.