Universal Basic Income and AI-Induced Job Scarcity

1329209391936196932

T-800 concept art by Victor Martinez

Mr. Charles Murray is floating an intriguing proposal. He suggests replacing social security, Medicare, Medicaid, and every other social assistance program with a universal basic income (UBI) of $13,000 a year. Murray, the author of The Bell Curve and propounder of spurious views on human group differences, is a controversial figure. But with his UBI proposal, he may be onto something.

Robots Took Our Jobs!

Looming just over the horizon is a world in which the proliferation of artificial intelligence (AI) and sophisticated automation stimulates colossal economic change. There are good reasons to believe an AI revolution will, over the coming decades, profoundly reshape labor markets, potentially producing unprecedented levels of jobs scarcity. Estimates of actual unemployment vary, but one Oxford study places the proportion of jobs threatened by advances in AI and automation at 47%. .

These, of course, are modern incarnations of the fears expressed by the Luddites, who over two centuries ago saw in technological progress the demise of their economic prospects. But as industrialization changed the economy, new jobs were created in unforeseen markets and swaths of textile workers were not condemned to perpetual destitution – instead, they got new jobs doing different things. Since then, fears of economic Armageddon have been quenched by a simple reminder that, for centuries, labor markets have consistently adapted to innovation. Consider the views of two advocates of this perspective, taken from a 2014 Pew Research canvas of expert opinions on AI and automation:

Historically, technology has created more jobs than it destroys and there is no reason to think otherwise in this case. Someone has to make and service all these advanced devices.

Vint Cherf, vice president, Google

In general, every wave of automation and computerization has increased productivity without depressing employment, and there is no reason to think the same will not be true this time. In particular, the new wave is likely to increase our personal or professional productivity (e.g. self-driving car) but not necessarily directly displace a job (e.g. chauffeur). While robots may displace some manual jobs, the impact should not be different than previous waves of automation in factories and elsewhere. On the other hand, someone will have to code and build the new tools, which will also likely lead to a new wave of innovations and jobs.

Michael Kende, economist

Unfortunately, the notion that this pattern of adaptation and perpetual job creation is an inborn component of markets is more of an article of faith than a statement of physical law. It is basically the hot-hand fallacy – i.e., the idea that a favorable pattern or trend will continue indefinitely – writ large. Cherf says that someone will need to build and maintain all our new gadgets. Sure, but there’s not reason to think AI and advanced robotics can’t fill that role. Likewise, Kende says someone will have to write the code that runs the AI – again, true, but also well within the potential wheelhouse of sufficiently advanced AI.

Not only do innovations in AI and automation have the potential to replace jobs in markets traditionally threatened by technological progress – manufacturing, unskilled labor, agriculture – but also a significant proportion of jobs involving service or the application of pure intellect. Cashiers are already being replaced by automatic check-out stands. People can order fast-food using Smartphone apps, reducing the need for workers to stand behind counters and interact directly with customers. Soon, self-driving cars (and the underlying technology) will eliminate jobs in long-hauling trucking, cab driving, bus driving, forklift operation, agricultural combine and tractor operation, and nearly every other form of terrestrial transportation of which you can imagine. Maybe the economy will adapt and new jobs will emerge. Yet it is difficult to imagine any job that can’t be replaced by sufficiently sophisticated AI and increasingly dexterous robotics.

Predictable Criticisms of UBI Initiatives

Opponents of UBI proposals offer resistance on a number of fronts beyond the traditional, pseudo-mystical, “The Economy will provide” mantra. Foremost among them is the hypothetical and value-laden argument that UBI will reduce incentives to work. Though there is little reason to think this will be the case, let’s grant the argument and accept that in a post-UBI world, fewer people will feel compelled to get off the couch and find gainful employment. This allows us to move on to the problem that emerges in its wake. Regardless of anyone’s ideological preferences or particular beliefs about self-sufficiency, hard work, and individualism, the naked fact remains that over the next half-century innovations in AI are going to obliterate millions of jobs across nearly every sector of the economy. Either we will come up with some kind of socially responsible means of addressing the problems those changes present or we won’t.

Opting for the socially responsible route is going to require some difficult choices and involve significant amounts of compromise across the ideological spectrum. The laissez-faire route, on the other hand, quite clearly spells disaster – boldfaced and underlined. It will involve abandoning a massive swath of the population to absolute economic disenfranchisement. The results of this are impossible to predict, but it’s hard to conjure a scenario in which they work out particularly well for anyone.

Perhaps the newly jobless will hive off into a separate society, scrambling for scraps and bartering for trinkets. Some might retreat to the hills, building a new economy on the primitive architecture of a pre-AI world. But if such fission happens, it will necessarily spell competition over resources between those fortunate enough to be on the right side of the AI revolution and those crowded out of work by inexpensive technologies. Almost inevitably, in will involve gobbling up the last remnants of the natural world as the jobless clutch at anything that might pass as arable land and scramble to secure energy resources.

Yet this presents a problem for those who remain in the AI economy. If such a large segment of the population drops out of the economy, it should mean a significant reduction in total purchasing power – fewer people buying fewer things, resulting in lower profits and deflated growth and the risk of wholesale economic collapse. Moreover, those marginalized by abject destitution will surely begin to feel isolation from the larger body of society and lose whatever faith they had left in the legitimacy of state authority. They will, in short, have few reasons to adhere to the laws that are meant to guarantee peace and stability in modern society.

Given this, it’s hardly surprising that Charles Murray has been thinking about UBI. Conveniently, it allows him to do some fancy intellectual gymnastics and posit it as a solution more closely aligned with his libertarian ethos than programs like social security and Medicare. More critically, it is also among the most practical alternative solutions to our impending problems. Murray’s specific proposal involves giving everyone 21 and older $13,000 a year. The first $3,000 is automatically allocated to purchasing insurance (whose costs should, presumably, go down when they replace all customer service representatives with AI). The next $10,000 is dispersed monthly, in chunks of roughly $833. If someone has a job, they still receive the full payout up to a salary of $30,000 a year, at which point their UBI allotment is gradually reduced to a floor of $6500 at a salary of $60k per year (people with healthy salaries still get a payout because in Murray’s vision, social security is gone and UBI has taken its place).

Problems with Murray’s Proposal

Murray’s proposal is not without its flaws. For one, it seems to give no attention to the needs of anyone under the age of 21. $13000 a year is probably enough to provide meager sustenance for a single individual, but seems woefully inadequate for anyone attempting to raise a child. Presumably, he would respond by pointing out that non-governmental volunteer organizations will fill this role. But this solution actually has the effect of illuminating a number of potential problems. Among them is the transactional nature of many non-governmental charity relationships.

A person raising children in a small town will likely have recourse to a limited number of charitable organizations for assistance and relief, which raises the questions: what if they exact a toll she cannot pay? The local charity offering the most pertinent resources might be a religious institution that demands a certain level of ritual observance and participation in exchange for charitable offerings. If the recipient is not a member of that religious sect, should she be forced to capitulate to their sectarian demands – attend weekly services, adhere to doctrine, undergo baptism? This, I think, would be unjust – it would boost the rolls of the local church by violating the dignity of the already downtrodden and disenfranchised. Frequently, private and sectarian solutions are by nature exclusionary, whereas public secular solutions can be constructed in such as to be necessarily inclusive.

There is also a problem of absolute cost and sustainability. And this is basically the same problem plaguing current social assistance programs, just expressed in a slightly different context. Murray wants to finance his UBI program by eliminating all welfare – that includes social assistance programs for the old, poor, and infirmed and profitable tax breaks for corporations. UBI, he argues, will be cheaper than the current system. This may well be true, though estimating the annual cost of UBI depends on a number of assumptions about unemployment and wages, making it difficult to check Murray’s math.

In any event, it’s not worth splitting hairs over cost estimates this early in the game. Everything here is purely speculative. Granting that UBI will prove cheaper than current programs, the fact remains that if birth rates continue to stay above replacement rates and AI steadily eats away at the number of available jobs, the costs of UBI will tend to increase even as the pool of taxable revenue decreases. A sustainable UBI program will thus entail either a dive into deeper debt or periodic tax increases on the gainfully employed. Few people are big fans of those things now and I can’t see many being big fans of them in the future.

So yeah, Murray’s UBI proposal (along with all UBI proposals of which I’m aware) is flawed. Which doesn’t offer grounds for an outright dismissal of the underlying ideas. Rather the opposite – they deserve careful consideration, a process which demands serious criticism and debate. The weaknesses of any serious response to the impending employment crisis will need to be uncovered under the lens of rigorous – and ruthless – bipartisan scrutiny. Better to spot potential systemic weaknesses in the planning phases, before they are manifest in the cold reality of people’s daily lives, than to allow them to unfold unexpectedly.

Looking Forward

Unsurprisingly, pages could be filled chasing the possible repercussions of any given solution. It only takes a little bit of time to recognize the take home point: there aren’t really any ideal options. If the widespread deployment of AI begins to precipitate increasing levels of irrecoverable job loss – and there are good reasons to think that it will do just that – it will demand a fundamental reworking of modern economic systems. The only question remaining is how fundamental? There are some people who advocate the complete dissolution of the money-based economy. This is almost certainly too far-fetched for the foreseeable future, but somewhat less fantastical than staunch market fundamentalists might suspect.

In comparison, UBI is a significantly more modest plan of action. The specific incarnation of UBI Murray is advocating has its problems. One might even go so far as to call it deeply flawed. But it is, I think, on the right track. At the very least, he is taking a serious look at the future and brainstorming about workable solutions to imminent problems.

Most prognostications share one thing in common: they usually turn out to be wrong. There are too many unforeseeable and unaccountable variables at work to come up with a solid prediction about the kinds of changes an AI revolution will induce. That it will eliminate some jobs seems likely, but it may be the case that social pressures and startup costs limit the scale or slow the pace of job loss (or both). But the very fact that advances in AI have the potential to radically alter the economy and eliminate increasing numbers of jobs makes the business of seriously considering prophylactic and remedial responses worthwhile. It’s the difference between being prepared to respond flexibly and being completely blindsided.

Maybe innovation in AI and automation will prove harmless, increasing productivity and opening unforeseen opportunities for gainful employment. This has tended to be the case with technological progress for the past couple of centuries, so why shouldn’t it continue to be the case in the future? But there is a certain level of risk to this kind of argument. It’s the thinking that guarantees the turkey is surprised when the hairless ape who has been feeding him for the past three years suddenly chops his head off.

FrameBreaking-1812

Luddites breaking a loom

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s