Some people believe that application of artificial intelligence will contribute in losing millions of people their jobs while others believe that a lot of people will be employed in this field.
I think it will decrease because at the implementation level, shortage of skills, the market encroaching by the capitalist monopolists, and being lagged by the poor industry sector which will market out due to incurring losses and being behind the competition from the mighty capitalists.
Did the invention of the hammer create more or less jobs?
Did the invention of the hammer create more or less murders?
What happened was just different uses of the hammer. If people chose to beat other people, or build homes were just different uses. Same with AI -- different uses will follow. It is difficult to know which those uses are, but some will be frightening and demand the development of defenses. I guess in total more and different jobs, and more and different disasters, will follow.
It eliminates many jobs and creates much less and in other fields. So, yes, the tendency will be that less and less humans will have work to do at all, which raises the question of how people will live, since the money that most, people use to pay for goods and services is what they are paid for doing their job.
On the other hand, the answer might be, like many other people, that, also, have money, without having to work for it. Many points need figuring out...
As a translator, I belong to a professional group that was target of the prediction that "in ten years, computers will be able to do all that crap, and then y'all gotta go and get real jobs for a change" before i started into it, even before i was born. But, lo and behold - we're still here, we have turned algorithms' insufficient efforts to catch up with us into a tool rather than a threat, and still anybody in their sane mind is well-advised to put all their eggs into the MT basket only if their texts are highly predictable in structure (weather report), or of entirely local relevance in content (way to station, food on menu), or if they couldn't possible care less about their audience.
The principles behind the most recent wave of A"I" text and image generators being the same as those behind MT (contextual net fishing, rather than straightforward parsing), the same limitations apply. [I could go into considerable detail at this point, but I'll better skip that for now...] Prague structuralism probably would describe these generators as conventionalization engines that are incapable of a single real actualization, and i cannot think of a better description.
So... be worried about your job only if you're doing superpredictable stuff, or things of very restricted relevance, or if you work for people who consider you a dispensable nuisance in the first place. Everybody else, prepare to be bored to tears with neverendingly inane simulations of human discourse that will keep growing in every nook and cranny until some realize that this has gotten seriously stale.
The part about this very recent concern about jobs-in-danger, moratorium etc pp that has some real entertainment value though... although translators have been described as targets for that for decades (see above), these concerns have become really loud (for all i can tell) only after a handful of focused articles and essays on, Might ChatGPT spell an end to tech and programming as we know it?
50 years ago while doing an engineering Masters there was much talk about 'self-driving' cars. There still is. It was obvious back then that risk and accident liability would stop this concept in its tracks. As is happening with the latest version. However, with an ageing population and folk dropping dead at the wheel, insurance companies will soon start to offer lower premiums for cars with systems that safely stop and park veiculos in an emergency. The motor trade will not develop it as it clashes with the image of their product, so we will have to wait for this to be good business for the insurance lobby.
AI is similar, it presents a vision of a Brave New World (the original line from the Tempest) and will end up a useful tool, especially in a world where data, texts, papers, etc., all increase exponentially.
However, it will increase the number of 'surplus people' (as definied by Hannah Arendt in Totalitarianism) to add to the huge numbers of the very old. An Unbrave New World (that has such people in it).
Warfare was changed in the recent Armenian Azerbaijan conflict with lurking drones hunting out their targets with AI. Russian tactics and tanks were swept away.
As with hammers, new stuff normally ends up being used in conflicts. These Brave New Wars with multiple forms of AI in use (hitting ultra fast missiles, etc.) are now too deadly to wage - as the long post war analysis of Ukraine will show.
Another use for both AI and self-driving vehicles that can already be seen (in Brazil) is in modern agriculture: huge machines covering the fields, checking the soil quality and adding lime and nutrients on a sq metre basis.