You may be unemployed when you grow up.
We have no idea what the job market will look like in 2050. It is a common belief that machine learning and robot technology will affect almost every field of work, from yogurt making to yoga instructors. There are opposing views only as to the nature of change and when it will occur. According to some, in a short period of ten to twenty years, billions of people will become dysfunctional to maintain the economic order. Some are of the opinion that even in the long run, automation will continue to create new business lines and provide prosperity for all.
So are we on the edge of a terrifying transformation, or are such predictions just another example of Luddite-like feelings based on unfounded foundations? It’s hard to say. Fears that automation would lead to mass unemployment date back to the 19th century, and until now these fears have not been seen to be realized.
People have two types of skills, physical and mental. In the past, machines were directly competing with humans in the field of physical skills, #people retaining their enormous mental advantage over machines. With the exception of this, when manual work in agriculture and industry became automated, new fields of work emerged to keep them functioning, and these required only the mental #skills of people: learning, analyzing, communicating, and above all understanding human emotions. But AI is increasingly at a level that surpasses humans in many of these skills, including understanding human emotions. We are unaware of the third area of activity outside of the physical and mental sphere where people can continue to retain an advantage.
It is extremely important to realize that the artificial intelligence revolution is not just about computers getting faster and smarter. This revolution is also fueled by advances in the sciences and social sciences. The better we understand the biochemical mechanisms that underpin human emotions, desires, and preferences, the more successful computers are in analyzing human behavior, predicting people’s preferences, and replacing human drivers, bankers, and lawyers.
This means that artificial intelligence can prevail on humans even in jobs that are supposed to require “intuition”. If you think that artificial intelligence will continue to compete with humans through mysterious hunches, this sounds impossible. But if AI is going to compete with neural networks in probability calculations and pattern recognition, this will sound less frightening.
Artificial intelligence can perform particularly well in tasks that require intuition about other people. Many branches of work, such as driving on a pedestrian street, lending money to foreigners, and bargaining for business understanding, require the ability to accurately assess other people’s feelings and desires. Is that boy going to jump on the road? Will the man in the suit opposite me take my money and shed? Is this lawyer serious about his threats or is he bluffing?
Artificial intelligence is not only meant to hack people and outperform them in skills that were hitherto considered human. In addition to that, they also have non-human skills. So the difference between artificial intelligence and human work is not just a matter of level, it is a matter of quality beyond that. Of particular importance are two of the inhuman skills that artificial intelligence has, that they can be linked and updated.
Since they are individuals, it is difficult to connect people and keep them all up to date. Computers, on the contrary, are not individuals, and it is very easy to include them within a single variable network. Therefore, what is at stake is not the replacement of millions of individual workers by millions of individual robots and computers, but rather the replacement of human individuals with an integrated network. So when evaluating automation, it would be wrong to compare a single human driver with a single autonomous car or a single human doctor with a single AI product. What we have to do is compare the skills of a group of people with those of an integrated network.
Such a thing has tremendous benefits for human society. AI physicians can provide billions of people with no healthcare, especially at present. The poor villagers of the underdeveloped countries can reach a much better health service than the richest people get in the most developed city hospitals today, thanks to the learning algorithms and biometric sensors through their smartphones.
From this point of view, it would be insane to avoid automation in transportation and healthcare fields just to keep people out of their jobs. After all, what we really need to protect is the people themselves, not their jobs. Drivers and doctors who are out of business will need to find something else to do.