I was recently asked to lead a workshop for Tech 2025 that interrogated the issues surrounding workforce automation and its growing association with universal basic income (UBI) programs. As someone who’s very much “in the bubble” on tech trends and more than a little biased towards an equitable economy without work (or at with least less work) it was refreshing to discuss these ideas with peers in New York who are bit more grounded than I am. That’s not to say that we didn’t end up discussing how to achieve a society modeled on Star Trek, because we certainly did that too.
In a way, the event could not have been timed better — we met just one week after the McKinsey Global Institute released a new report on workforce automation, and I began my workshop by discussing some of the findings. Perhaps most shocking, the report indicates that 49% of today’s work could be automated with currently existing technology. Such technical capacities are not the only factor, however. The report suggests that obstacles like labor market dynamics, regulations, social norms, and other economic trends could delay the automation of various tasks until 2055 — give or take 20 years. (It’s important to note that the low end of this estimate, 2035, puts the McKinsey projection in the same ballpark of the often criticized Oxford Study by Frey & Osborne).
Keeping in mind that McKinsey’s study is looking at tasks within jobs, researchers predicted that only 5% of all jobs were likely to have their human components automated away completely. However, the skills most likely to fall into obsolescence are those possessed by middle income workers — accountants and radiologists, data analysts and legal aides. As the middle class contracts amidst a wide array of pressures — automation only one among them — the end result is the economic polarization that David Autor warns about. I’ll bluntly characterize that polarization as the stratification between a ruling class of those who possess capital and a servant class of those who do not, with few opportunities for social mobility in between. To me, this is the real threat of more automation without fundamental changes to our economic system: not a jobless future, but a future defined by a large glut of precarious, low-paid work.
In the context of these trends in automation, one key issue area we focused on is why Sam Altman, Elon Musk, and other Silicon Valley elite support a basic income program that presumably would require higher taxes, likely to disproportionately fall on people like them. On the one hand, we can be overly idealistic and presume that these people are genuine humanitarians that are excited about the possibility of basic income as a “social vaccine.” Or we could perhaps conclude that basic income as “venture capital for the people” fits the ethos of Silicon Valley. But being more cynical we might also suggest that without some means of financially stimulating a growing class of economically marginalized people, there will be no one to purchase their new gadgets, no one to use their new apps; therefore basic income is a modest program of wealth redistribution focused about sustaining a certain level of consumption. Or further still, we may say that without some subsistence level of redistribution, the titans of the new “gig-economy” will be unable to find precarious workers to accept lower and lower wages.
In the end though, it’s worth reminding everyone that when Thomas Paine talked about a tax on landowners that would be redistributed across society, he wasn’t concerned about robots. Martin Luther King, Jr. wasn’t concerned about the implications of machine learning and algorithms on middle income earners when he argued for a guaranteed minimum income. Basic income makes great sense if you’re looking to change incentive structures that surround various forms of labor or create an income floor for marginalized and underserved peoples. However, the notion that the program can solve the dilemma of a dwindling amount of work or, more likely a dwindling amount of dignified work, because of technological advancement is a very recent and specious invention.
And it seems that my guests at the workshop largely agreed. After an initial presentation, I divided the room into four groups (precariat, middle class, business elites, and policymakers) to explore what the automation dilemma looks like from each perspective. Then new groups with at least one representative from each of the initial four met to negotiate a new deal. One group agreed on a guaranteed minimum income, another group agreed on a negative income tax system. But in the end, most of the attendees were suspicious of a basic income policy. Some were concerned about the rise of a “free-loader” class, others weren’t convinced that automation was the paradigm-shifting force that I portrayed it to be. Rather than grasping at UBI as a transitional step to a post-work future, many were actually interested in jobs programs like mandatory national service or job guarantees. Rather than a universal “handout,” many were interested in strengthening the social safety net, in particular with universal healthcare and universal education programs.
At the risk of seeming a rude host, I would like to take a critical approach to the solutions that the workshop produced. First, there was a reluctance on their part to view automation as a fundamental paradigm shift to the way we organize work in society. Typically, we might think of a wealthy class of capital owners that has to bargain with labor (hopefully unionized!) in order to get that capital operating at an efficient level. Increased automation means that more and more the role of capital is expanding and crowding out labor. Assuming any rate of improvement, firms will eventually be able to automate any task that humans currently do, whether it takes until 2035, 2055, or beyond depends on a wide range of factors. It’s very difficult for labor to unify and press their demands when they simply aren’t as essential to production or even when fewer workers are needed. As I’ve argued elsewhere, we need institutions and policies for baking democracy into our economic activity.
Furthermore, while I believe in a universal guarantee of higher education, I don’t think it’s a safe bet for adapting to the robot revolution. On the one hand, education plays an important social function beyond transforming pupils into economically productive units. Education is important for education’s sake, and I would hesitate to relegate it simply to what markets deem profitable. On the other hand, higher learning isn’t proven to safeguard jobs from the capacities of technological advancement. In a Q&A for Future Left, Andy Stern pointed out that from accountants to lawyers to adjunct professors there is no guarantee that higher learning will equate to job security. The way that jobs will transform in the next five years, let alone the next ten years, due to tech advances including machine learning, artificial intelligence, advanced robotics, augmented reality, additive manufacturing, et cetera ad infinitum, means that the paradigm of learning skills and trades when you’re young so that you can use them for the rest of your life over a long career is simply coming to an end.
When it comes to planning out policy responses to a more automated future, there are many suggestions about the best approach. Universal basic income, job guarantee programs, shorter-work weeks, strengthening the social safety net, distributed ownership schemes, have all been suggested in one form or another by a variety of different sources. My feeling is that the best approach is likely all of the above, with plenty of experimentation and a willingness to boldly upend assumptions and paradigms as we attempt to deliberately and democratically design the future.