AI is great for pattern recognition associated with repeatable events.
However, in dealing with “problems” where things are not repeatable human intuition and experience will still be valued.
Another problem with AI has to do with “scale.” For example, let’s talk about electrical codes and design standards. While an AI algorithm can be designed for say troubleshooting the National Electrical Code, there are slight differences in the Electrical Codes of different States. Furthermore, there are differences in the code adopted by various cities. Finally, many companies and government agencies have their own “standards” for electrical design that differ from City, State, and national codes and standards. These different codes and standards have come into being because of problems and disasters that happened in the past. The question is how much will it cost to modify the AI software for all the really small scale “systems” and is it economically worth the savings, especially when those codes and standards are changed every few years?
The same thing is true in terms of troubleshooting. If something is repeatable and predictable, AI is great.......as long as it is fed the right data that describes the problem. But who decides what data to take and what data to not take to feed to the AI model. If you say that the AI model will tell you, then for unique problems or multiple problems, the AI might be fed bad data that produces wrong results. Someone still needs to check that the AI solution actually worked.
Yes, AI will take lots of jobs, but this is not new to history. Look at the industrial manufacturing process. There is artisan hand assemble for customer or small one off things. There are assembly lines with human assembly workers. There is robotic assembly for larger scale production. Then there is “hard automation” for really huge production numbers.
The jobs that AI will take is analogous the jobs lost to automation. There is a scale at which jobs will not be lost and a scale at which AI will be the only reasonable choice.
AI will likely become part of most future jobs, but I think that there will still be highly paid jobs out there for those that have a huge variety of skill sets.
What about Learing Specialist?
Prior to the Industrial Age, a low-I.Q. individual was still able to "get a job," say, working a treadmill to irrigate a field - i.e., performing a task that could have just as easily been performed by an animal.
In the Modern Age, it is nigh impossible to find gainful employment for such intellectually limited individuals. There are, after all, only so many cushy govt. jobs to go around! [Heh-heh!]
It is, of course, still possible to create "make-work" for them, as long as they are under supervision; I'm not disputing the social benefit and positive humanitarian effect of maintaining such programs to give low-I.Q. people the "illusion" of gainful employment, but I assume that we are agreed that it would be an obscenity if such jobs became the norm for vast swathes of our population.
I agree with your thesis statement: Even after the advent of A.I., there will still be "jobs" out there for highly qualified individuals. But your statement lacks the all-important quantifier: for the increasingly tiny sliver of highly qualified individuals.
How can a society with any self-respect continue when only, say, 10% of its members are performing actual work - and the rest are, essentially, useless "drones?"
The loss of human agency, the social stratification caused by rampant automation, and the psychological impact of being rendered obsolete by technology would have an immense impact upon the human species. (All discussed quite accurately by the prescient science-fiction novel Player Piano by Kurt Vonnegut Jr. in 1952!)
Regards,