These days, it seems that the discussion on how AI is going to disrupt the vast majority of industries in just a few short years rages everywhere. According to PitchBook, in 2017 VCs have poured more than $10.8 billion into AI & machine learning companies, while the incumbents have spent over $20 billion on AI-related acquisitions; according to Bloomberg, the mentions of AI and machine learning on earnings calls of public companies have soared 7-fold since 2015; and just this week, The Economist published a series of articles, framed as Special Report, on the topic.
In today's context, AI typically refers to machine learning, rather than any kind of attempt to create general intelligence. That, however, doesn't change the fact that the current technology has clearly moved past the point when it was of limited use to non-tech companies, and is now beginning to disrupt a large number of industries, including the ones that weren't particularly tech-savvy in the past. To quote McKinsey Global Institute's "Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation" report:
"We estimate that between 400 million and 800 million individuals could be displaced by automation and need to find new jobs by 2030 around the world, based on our midpoint and earliest (that is, the most rapid) automation adoption scenarios. New jobs will be available, based on our scenarios of future labor demand and the net impact of automation, as described in the next section. However, people will need to find their way into these jobs. Of the total displaced, 75 million to 375 million may need to switch occupational categories and learn new skills, under our midpoint and earliest automation adoption scenarios."
To be fair, McKinsey also states that less than 5% of all occupations consist entirely of activities that can be fully automated. Still, here's another valuable quote from the report:
"In about 60 percent of occupations, at least one-third of the constituent activities could be automated, implying substantial workplace transformations and changes for all workers."
Overall, there seems to be little doubt today that even with the current level of technology, the global workforce is about to enter a very volatile period that would require large numbers of people to learn new skills or be altogether retrained, or else risk losing their jobs, and face difficulties finding new employment.
The peculiar nature of disruptive technology adoption
I would also argue that while tech industry, as well as the broader society, have often been overly optimistic when trying to forecast how soon certain revolutionary advances in technology were to happen (heck, in their 1955 proposal the fathers of AI, which included Marvin Minsky, John McCarthy and others, outlined their belief to be able to make significant progress towards developing a machine with general intelligence in a single summer), once the core new technology became available, even the most daring forecasts for adoption rates often turned out to be too conservative.
This is especially true in cases when technology in question was impactful enough, and its nature allowed for the formation of an ecosystem around it — in which case, in just a few years, there were hundreds of thousands of stakeholders involved coming up with new creative ways to benefit from the advantages brought by the new tech.
With AI, or rather, with machine learning (in this case, the distinction is quite important), while the underlying technology is still evolving and will continue to do so, it's already good enough for a wide variety of applications, which prompted a rapid rise in the number of tech companies, startups, consultancies and independent developers involved in the space — today we already have a vast ecosystem around AI, with the ever-growing number of stakeholders involved, and it can only be expected to grow larger in the next few years.
Rethinking the safety nets
What that means is that even the most daring forecasts produced by McKinsey or anyone else might still underestimate the change that's coming. And if that turns out to be true, figuring out how to help all the people who are going to be displaced, becomes of utmost importance, as the society will need to find ways to support those people through periods of unemployment, provide them with the training that would be effective in bringing them back to the workforce (the current government-run retraining programs, while costing a lot to the taxpayers, often turn out to be painfully ineffective, at least in the U.S.), and, ultimately, take care of those who for various reasons can't get back to the workforce, while doing all of the above on an unprecedented scale.
This calls for the creation of robust safety nets for people, while also making sure that it doesn't stifle economic growth: while the safety nets of some European countries are great for their citizens, they also place undue burden on the employers, and incentivize both the mature companies and the startups to move their business to other places, if possible (and in the increasingly global and interconnected world, it is indeed becoming possible to do that more and more frequently).
At the first glance, there is a paradox here: the safety net is becoming increasingly important, but if a robust safety net stands to hurt the economic growth, then there'll be less jobs going on, in turn making the safety net even more essential, and more costly to provide. This paradox, in turn, brings the ultimate question: why are our safety nets designed with the assumption that it's the end goal for people to have a formal full-time job? Note that this is the case for most developed countries, including the U.S.: while it might be easier to fire people in the States compared to many European countries, the system is still designed to incentivize people to seek full-time employment, in some ways even more so than in Europe.
If you think about it, it doesn't seem to make much sense to force people to look for full-time employment above everything else, or to force the employers to make long-term commitments to their employees, or to bear most of the burden associated with their safety nets, in a world that is increasingly global and going through rapid changes at an accelerating pace. Wouldn't it be better if at least most of the safety net would come from the state, while the employers would be incentivized to optimize for efficiency and growth, bringing in people (and firing them) as needed?
This added flexibility for the employers doesn't need to be free either: it's no secret that the corporate taxation is dysfunctional, but it's hard to fix it without offering a decent reason for the companies to play nice (instead of moving the profit center to Ireland) and comply, and the added flexibility on managing their workforce can potentially be a powerful incentive (especially in the HQ markets, where workforce constitutes a significant expense, and can't be easily moved elsewhere). For the businesses, that would mean that they are still being asked to pay their fair share, but at least they won't have to make upfront and long-term commitments that can often have perilous consequences in the changing markets. That remains particularly true for the smaller companies.
Would such a world be more volatile for regular people? Alas, it most likely will be. But it also stands to reason that in a world where your health insurance isn't tied to your employer but is instead provided by the state no matter what and where you have the opportunity to go back to the school as needed, without having to worry about the cost, people would be much more daring to pursue the career options that are best for them long-term.
The final piece: UBI
There is still one component missing, of course. If there is nothing preventing your employer from firing you without much notice, the safety net has to include some mechanism to account for that, and, most likely, it has to be more robust than the currently available programs, which brings the conversation to the concept of UBI, or universal basic income.
Now, that's an incredibly broad topic, and the one that has been in discussion for decades, if not longer (for example, few people know that the U.S. actually conducted a number of experiments on negative taxation way back in 1960s, and even almost got to implement a form of basic income). Also, basic income doesn't stand for one particular idea, but rather includes a range of concepts, from offering everyone the same lump sum regardless of their income or wealth, to ideas of negative taxation that would help to create an income floor for everyone, to proposals that are more limited in scope, but might still play a valuable role in helping to eliminate poverty and providing safety net for people.
The most realistic concept I've seen so far, and the one I like the most, is described in the recently released book called "Fair Shot: Rethinking Inequality and How We Earn", written by a Facebook co-founder Chris Hughes. I'd highly recommend reading the book to anyone interested in the topic, but in short, the idea is to supplement the earnings of every household with the annual income of $50,000 or less, with additional $500/month per working adult (less, if the income is close to $50,000), building on top of existing EITC program, and to pay for this program by eliminating preferential tax treatment for capital gains, and imposing additional taxes on those who earn $250,000 or more per year.
While this idea is less daring that the some of the more sweeping concepts of UBI, it has several extremely interesting components to it. First, it's much less expensive than some of the other UBI proposals, which in theory means that it's possible to implement it even today. Second, unlike the calls to provide basic income to everyone regardless of their wealth or whether they are working or not, Chris proposes to provide this supplementary income to working adults with relatively low earnings, but to use a much broader definition of work that the one currently used in EITC: the idea is to count as work any kinds of paid gigs (e.g. working for Uber, TaskRabbit and the likes), as well as to count homemaking and studying as work. That way, people would remain incentivized to engage in productive activities, but wouldn't be limited in what they could do as much as they are now (although, interestingly, the vast majority of UBI experiments actually provide evidence that people receiving it continue to work, and even work more, instead of withdrawing from the workforce, so this concern is artificial in nature to begin with). Third, while $500/month won't be enough to support someone who has no other income, its value shouldn't be underestimated: the studies show that even small amounts of cash can help people get by during the hardest periods and optimize their careers for longer term.
The path ahead
Even with the UBI in some form, the guaranteed health insurance and the access to free education, people wouldn't exactly get to enjoy their lives without having to worry about work: the goal, at least for now, should be to provide safety net for the periods of turmoil and incentivize people to pursue riskier and more rewarding opportunities career-wise, rather than eliminate the need to worry about finding employment altogether. Still, having this safety net would mean a great deal for someone whose job has been eliminated by automation and who now has trouble finding work, or who needs to go back to college to get retrained, or simply wants to quit her less than inspiring job and try to launch a business.
The change brought by the globalization and automation is inevitable, and so, most places would have to find a way to adapt to it, one way or the other. Right now, places like Netherlands or Nordic countries already have well-developed safety nets, but often represent a challenging environment for the new businesses to grow in, while other places (e.g. the U.S.) can be much more business-friendly, but don't offer all the necessary protections to support people who find themselves worse off than before. What remains to be seen is what path each of those countries would choose to pursue going on, and how it would play out for them over the next 10-20 years.