Why don’t we just give everyone $20,000, no strings attached?
That, in essence, is the policy known as Universal Basic Income, or UBI. It has gained popularity on the right and the left because of its simplicity. It would replace the basic income already provided by complex welfare and unemployment insurance plans, instead handing an above-poverty amount to every adult regardless of income or employment status.
UBI appeared in the news recently when Ontario Premier Doug Ford abruptly cancelled a provincial pilot study that was giving slightly smaller amounts to about 4,000 people. The Premier was wrong to cancel the study; governments should never be enemies of knowledge.
But just because it’s hated by small-minded people doesn’t mean UBI is a viable policy, or even a desirable one.
UBI has moved from the fringes to the mainstream because some believe it may soon be an emergency measure. They believe we will face mass permanent unemployment because of emerging artificial intelligence (AI) technologies rendering many jobs obsolete, including a good number of skilled trades and professions.
Indeed, a number of recent studies project that half of all currently existing jobs will vanish by mid-century because of AI. Tax lawyers, family doctors, maybe dental hygienists will go the way of telephone operators and video-store clerks. Under this logic, the only way to keep people alive – and purchasing the products made by those robots – is to give them a state-provided income.
Whatever its intrinsic merits, a UBI program would be a political non-starter if we didn’t face huge-scale joblessness. It would entail a substantial tax increase on the employed (offset only partly by that free $20,000). If unemployment remains in the single digits, it won’t happen.
Yet the most informed and forward-looking economists and AI experts feel that even the most cogent AI machine-brains will not eliminate employment. They may well kill half of all current jobs – but they will provoke the creation of at least as many as-yet-unknown jobs.
“There will always be enough jobs – just the nature of them will change,” says Marcel Fratzscher, the renowned Frankfurt economist who runs the German Institute for Economic Research. “There will be a thinning out of the middle class in most Western countries, because those middle-class jobs are the easiest to replace through technological change, and that is already happening.”
Branko Milanovic, the City University of New York economics professor who became an authority on global income inequality during his years as the World Bank’s chief economist, agrees. “I am quite optimistic on jobs under AI, but I’m very pessimistic on equality,” he says. “The number of jobs will probably expand – but it’s an epistemological problem, because we can’t know what kind of new jobs will be created and we don’t know which ones will disappear.”
Kai-Fu Lee, the Chinese AI authority, has a similar understanding: Smart machines will create at least as many jobs as they destroy, regardless of how smart they get. And AI’s efficiency – the only reason we use it – will expand production, and thus employment.
Consider your TV. A few decades ago, making a television involved hundreds of hours of human labour, so sets were once-a-decade purchases. Today, only a couple hours labour, at most, go into making a TV. Yet far more people have jobs making TVs today than ever before. Why? Because automation made the price of TVs plummet to the point that most of the world can afford at least one every couple years. It may take one-twentieth as many people to make a TV, but as a consequence there are 40 times more TVs being bought, so employment has at least doubled.
There’s no reason this effect won’t extend into services and professions: Architecture could go the way of TV-making – a $30,000 professional contract replaced by a $300 intelligent home-designing app, with 10 times more people employed as a result, but at considerably lower incomes.
As a result, many of those economists and AI experts feel that some form of more flexible state support may indeed be needed – as part of a larger plan to prevent a grossly two-track society with half of working people trapped and dependent.
But a universal income, they feel, would be not a solution but a cop-out.
““It washes the hands of government of responsibility for the problem‚" Mr. Fratzscher says. "A universal income just says, ‘Here’s some money, go off in a corner and stop complaining.’ A job is a source of self-worth. Buying people out of the labour market does not create an equal society, it just hides the problem.”
If machines get smarter, we’ll need intelligent policies – not ones that simply pay people to disappear from the human community.