In other news, Bloomberg is has now reached a new low for click-bait technopanic. Do they think the world will run out of low-skilled jobs because AI learns how to perform current ones? History doesn't back this up, as things like the industrial revolution, or automated farming, or etc etc etc didn't cause low-skilled jobs to disappear, it just changed what the exact nature of those jobs were. More and more i dislike seeing Bloomberg on here, they're "tech scare" approach to journalism is inaccurate and annoying, and it's causing a lot of unnecessary negativity towards new technologies.
> Do they think the world will run out of low-skilled jobs because AI learns how to perform current ones? History doesn't back this up,
I don’t have a well considered opinion about AI’s impact on jobs, but AI isn’t exactly like the industrial revolution, automated farming, or other historical upheavals. So I don’t think you can reliably conclude that it will turn out just like them.
This is a Bloomberg Opinion and written by Dr. Kai-fu Lee, former head of Google China, and former SVP at Microsoft and Apple. He was previously a world-class speech recognition/machine learning researcher at CMU[1] And at the bottom of the article, it reads:
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Say a factory is fully automated. What low skill jobs have been created? This isn't the same as the invention of the spinning jenny. Unreasoned panic is just as foolish as assuming that things will just sort themselves out for the best.
Some blue collar jobs that will be created include: maintaining the factory robots, producing the robots that run the factory, installing the robots that run the factory, and retrofitting old factories to work with new automated systems, lubing up the factory robots and replacing their oil and filters. Not to mention entire industries which will crop up - millions of specialized engineers, automated factory focused marketing firms, new government sectors for people to write regulation for automated robots, etc.
The world changes, but we never run out of stuff to do.
It's troubling because when I worked there as a developer their journalists seemed to be really on their game. I loved the UX of their site, and the articles seemed thorough, well written, and always relevant.
Yea i feel you, I've always read a lot of Bloomberg but, the last few years I keep seeing articles like this that are speculative but scaremongering. It's not a great look for them, in my opinion.
I dont think this article has much merit for a few reasons. The two category of Jobs that are outsourced to India or other developing world are Either service oriented or Manufacturing oriented.
For Instance The service oriented jobs in India like Customer support and Offshore Software development cannot be yet replaced by AI. Customer Support is trouble shooting driven and Software Development is innovation driven, the latter would take decades to be replaced by AI, at which point all the Jobs would have been anyway replaced by AI. The first category though could be replaced in the next decade or so.
The automation in manufacturing can replace some redundant jobs but not all of them as evidenced by Teslas Manufacturing fiasco where they tried to automate everything leading upto a delay, as humans are more dexterous and can do some tasks much better than Robots.
So I dont see how AI has reached the point which this article claims can replace human labour yet.
> For Instance The service oriented jobs in India like Customer support and Offshore Software development cannot be yet replaced by AI
Customer support is troubleshooting driven ? Heh. You must not have used it for a while.
But aside from that. Let's say that it is. That essentially consists of 2 problems:
1) voice recognition
2) query answering based on freeform text
(1) is solved, (2) is proving a little tougher, but pretty nice progress is being made.
> The automation in manufacturing can replace some redundant jobs but not all of them
Sure, at this point AI is slowly becoming a force multiplier for manufacturing. What a factory can do with one employee will steadily go up.
The issue is that currently this multiplier will kill a LOT of jobs. Let's say an employee productivity doubles (you only need 1 employee to do what 2 used to do), how many jobs does that destroy ?
About a billion.
Granted, it will take 15-30 years for that to actually happen, but there'll be an initial spike (let's say rapidly dropping 100 million jobs) and a slower but steadier increase over time until we hit that number.
Of course over 30 years the multiplier is not going to be 2, but hundreds. So another 30 years after that, manufacturing will have 2-10% of the population employed, no more, like agriculture today.
This is no surprise. This is the low hanging fruit. And economies that are already quite advanced (have outsourced the more mechanical work to developing countries) are at less risk.
Perhaps a silly example, but imagine people whose job it was to answer arithmetic queries: "What is 47 + 1.5?" Any modern human who couldn't be bothered with doing that in their head would use a readily available calculator, even if that was just an address bar in their browser.
Clearly, the arithmetic people would lose their jobs to technology. There's no creativity to their work; it is easily programmed.
This applies to any job. If your job is to vacuum a floor, then eventually there will be some version of an iRobot that can do it just as well or better.
Humans have immense potential, depending on how you choose to measure it. Maybe not all that potential is marketable, but most of it is currently not outsource-able to AI or robotics. Eventually, in the internet-connected world, if you want to remain relevant as a human, you'll need to exercise some of your talents that cannot be reasonably approximated by a computer. Obviously you'll stop playing chess for money... and you'll stop doing other tasks that robots or specialized hardware can do.
Humans were once the only means of labor; even animals needed a human attendant – and the cost of construction was the number of capable adults to help build economies.
Industrialization, and especially the use of oil and electricity to do work, decoupled man-power. Computers and the internet have decoupled man-hours, and now we see economies of scale, and an information economy. With AI, we may see even less dependence on human labor.
What will this do to the human population? With low skill labor made irrelevant, how will those people contribute and receive money from an increasingly gated economy?
Interestingly, food also benefits from economies of scale and automation. Cheaper food leads to a larger population. Larger population means more people contributing to the economy, but as we've already seen in farming this last hundred years, fewer and fewer people of whom the economy pays wages to.
It's a strange balance. I do not know what it all means.
I still don't know AI or robots capable of doing work of electricians, house constructors, shoe repairing, car repairing and so on. There's a class of job too complicated for a robot at the current state of art (say, for next 50-70 years at least), because humans do it better and even in awful conditions.
AI can't cope with bad weather or too complicated problems.
See how the "autonomous cars" drive well on sunny and dry days against windy or snowy days
There are easier problems that still cannot be solved by robots. The biggest milestone will probably something completely trivial like newspaper delivery. It's repetitive routine work with the added complexity that every mailbox is completely different.
The underlying assumption in the article is that even after the AI-Automation revolution people will still want jobs. A good chunk of the population might just be happy when their basic needs of food, shelter, clothing, entertainment, etc are met by automated factories and services. It is possible to create this supply meets demand world at zero recurring costs.
But then those people have no leverage to barter with and save them from oppression.
I worry that the ai powered future looks like a country with the resource curse - only those with control of the means of production have any political power.
How do we ensure people continue to have any real political power without their underlying economic power to enforce it?
