If you can’t completely automate something, there are by definition, still useful things to learn.
Outsourcing the advancement of low level work to automated work puts a business’s competitive edge in jeopardy.
this is not a black-and-white thing, outsourcing makes sense in many cases and doesn't in others. the automotive industry does it for quasi-standardized components. (and there's a competitive market, for airplane parts, not so much it seems.)
and on top of all this for things that have inherent safety critical aspects there is no real difference between outsourcing or doing it in-house, there's a need for controls (organizational, quality, independent testing lab with random sampling).
the fucked up process of gutting companies for short-term profits happened just as well for non-outsourced departments too.
as industries mature, companies grow and became organizationally more complex training and advancement itself becomes a challenge. (and many companies just put their heads in the proverbial sand, and hire clueless recruiters - or outsource it - and end up with paper tigers, who have absolutely no idea WTF is going on in the company, but quarterly reports look amazing.)
...
finally, if market forces push for short-term thinking (and simple, wrong, but fancy 'solutions') that's what we'll end up. and if society and politics also has the same mindset it's not surprising that industry will mirror it. (and to overuse the overused cliché gestures wildly )
I think it is fair to expect that everything will get automated in an economically advantageous way.
And if the solution is novel & challenging, it is likely to provide valuable unexpected side benefits & capabilities.
The automation question is just: when, how, and by who.