The idea of universal basic income (UBI) is not a new one. Its origins are hard to trace, but thinkers have been proposing that governments might pay all their citizens (including the wealthy ones) a standard wage since at least the 17th century.
The concept has reared its head over and over again in the intervening years; in one of the oddest twists in the story, fiscally conservative US President Richard Nixon almost introduced a guaranteed income for all Americans in 1969. Whether because of cost, marketing, or politics, though, no attempt to introduce basic income at scale has ever succeeded in a developed country.
However, the rise of artificial intelligence (AI) has given us a new, potentially viable blueprint for such a system. Workers across industries are worrying that robots might soon steal their jobs; if this does happen, could governments step in and start paying their wages?
Universal Basic Income in Practice
Various pilot schemes around the world have introduced basic income, but there have been significant differences between their approaches.
A Finnish experiment that ran in 2017 and 2018 provided 2,000 randomly selected unemployed people with €560 ($590 USD) a month, with no reduction or removal of the benefit if the subjects got a job. In 2011, Iran started paying almost all its citizens a monthly stipend of around $45 USD, simultaneously phasing out government subsidies for essential items like bread and gas.
The Mincome experiment ran in Manitoba, Canada from 1974 to 1979. Participating families in the city of Winnipeg and the small town of Dauphin received payments ranging from $13,000 to $22,000 (USD) annually in today’s money. The Mincome scheme involved partial payment reductions in line with changes in working income.
It’s worth noting that these schemes fall short of the true ideal of universal basic income, under which everyone, be they prosperous or penniless, would receive a living wage. Though this might sound frivolous, it has its advantages.
Crucially, full UBI would remove active incentives to stay out of the workforce. Whether you decided to work three jobs or no job at all, you’d keep receiving the same amount of money from the government every month. It would also solve many of the long-standing, expensive headaches around social welfare; with an unconditional income, no one would have a reason to exaggerate the symptoms of an illness or to lie about their single-parent status.
Though more large-scale pilot schemes would be helpful, the evidence we do have suggests that basic income might be a more effective way to distribute resources than current welfare frameworks. Studies have associated the implementation of universal basic income with positive trends in terms of employment, poverty, education, family unity, and personal wellbeing.
Though there are significant cost benefits to UBI (such as from doing away with social welfare enforcement and bureaucracy) the system’s price tag has always been its biggest issue. In the United States, for example, a monthly UBI payment of $1,500 to every adult would set the federal government back around $3.9 trillion a year, which is well over half what it spent overall in 2022. While the fall-off in welfare spending would pick up some of the slack here, Uncle Sam would still have to find a way to pay for UBI alongside the military, transport infrastructure, law enforcement, public healthcare, and other projects.
Now, though, with artificial intelligence seemingly poised to take over a lot of human work, that stumbling block could get a lot smaller.
The Rise of Artificial Intelligence
Since late last year, artificial intelligence has emerged from the realm of obscurity to become a major point of conversation for just about everyone, and those discussions are increasingly starting to center around concerns about white-collar work. It’s not hard to see why people are worried; the most current iteration of OpenAI’s ChatGPT, GPT-4, can pass American bar exams, code video games from scratch, and explain why jokes are funny.
A now-infamous April 2023 report from Goldman Sachs estimated that 300 million jobs worldwide could be at risk from automation in coming years. Workers in two-thirds of job types in the US could be affected, the report also claimed.
However, not everyone is bullish on the idea that human work is about to fade into memory. Roman Borissov, head of digital marketing agency SEOBRO, says that while his company has started using AI for “data analysis and trend forecasting,” it has not lowered its spending on freelance content writers (a group with particularly acute worries about artificial intelligence).
“While AI can create data-driven content efficiently,” Borissov states, “there are nuances, creativity, and a level of emotional understanding that only human writers can effectively deliver. As such, we have kept our spending on freelance writers, but we've refocused their efforts on creating more complex, in-depth content, where their skills genuinely shine.”
This idea that artificial intelligence will complement human endeavor, rather than replace it, is a common refrain among AI bulls. Chirpy entrepreneur Gary Vaynerchuk (more commonly known as GaryVee) is one such advocate; when concerned social media commenters ask him about job security, he blithely notes that 19th-century farm workers had the same worries when we invented the tractor.
My personal experience of the knowledge economy has made me skeptical of this type of optimism.
In my early days as a freelance writer, I did a lot of what is known as “content mill” work. I wrote long-form articles about everything from HVAC installation to vampire facials, published with the intention of boosting the search engine rankings of clients. The writing required very little prior subject knowledge or primary research; I just had to skim through existing content on the topic in question and regurgitate something semi-original.
One of the agencies I did this type of work for completed hundreds of these orders every day at the time, paying its team of writers and other content workers somewhere between $250,000 and $500,000 on a busy month. Based on the abilities of ChatGPT and the number of job opportunities for roles of this type at the moment, I am fairly certain that this work now requires very little human input.
While this work may be a short-term casualty of a shifting economy, it’s impossible to ignore the fact that AI is already costing people their jobs, and that it has the potential to do so across industries. Experts believe roles like coding, data analysis, advertising, market research, graphic design, customer service, and even teaching could be under threat from automation.
Even the rich and famous might not be safe. Earlier this year, Hollywood union SAG-AFTRA went on strike, citing artificial intelligence as one of its key issues. Specifically, “performers need the protection of [their] images and performances to prevent replacement of human performances by artificial intelligence technology.”
So, businesses clearly intend to use AI to widen their profit margins and limit their spending on salaries and benefits, and at least some of them seem likely to succeed. Could governments address the obvious potential problems here by restructuring the tax system and introducing some form of basic income?
The Road to Universal Basic Income
Otto Lehto, a philosopher and political economist at New York University, is optimistic about the future of basic income, but unsure how great a bearing the rise of AI will have. He says he hasn’t heard of any government undertaking a guaranteed income initiative as a direct response to recent developments in artificial intelligence, even though automation has been a factor in discussions around the topic since at least the 1960s.
The theory is that, if machines take over human work on a large scale, the organizations in charge of the machines will be able to drastically cut costs while maintaining or increasing their revenues. Governments could impose higher taxes on the resulting corporate profits, thereby filling the UBI money pot.
So, is the idea that governments are not at least considering some form of basic income program as a response to AI a worrying one?
Perhaps not. Lehto says he is “skeptical” of the idea that “automation should be the primary reason” for guaranteed income, explaining that “technological unemployment” may end up being a “transitory issue,” rather than a permanent one.
So, AI and universal basic income may not be joined at the hip. But, if ChatGPT isn’t going to pay the bills, who is?
Lehto believes many countries already have the money to establish a worthwhile basic income system. This would require a massive overhaul of our outdated social welfare frameworks and the dismantling of the “heavy bureaucracy” that exists within them. But, if we can achieve that, basic income provides the best available answer to the pressing questions around issues like retirement funding that are popping up across the Western world, Lehto thinks.
He says he would prefer to see funding for UBI come from additional taxes on “natural resources or land” rather than income. Such an approach could help to address the wealth disparities that have increasingly arisen within developed nations and mitigate the risks basic income poses to labor-force participation.
Lehto concedes that the “biggest concern” around basic income is the possibility it will disincentivize work, but states that this could be addressed by paying people the right amount. “It’s especially a concern for those models that are either too high (that is, they’re set at a level that leaves people with very little extra incentive to top up their income). Regarding the level issue, if you keep it at a sufficiently low level, you still incentivize people to find other sources of income. Some people will choose not to do so, but that’s perfectly fine; it’s a matter of statistical averages.”
He says most basic income experiments have shown that these models do not disincentivize work to a greater extent than existing welfare programs, and that no experiment has resulted in a significant drop-off in labor force participation.
Lehto points out that labor unions pose as big a threat to UBI as corporations and fiscally conservative campaigners. He explains that “interest groups” like these have backed traditional welfare schemes too strongly over the years to support a program that would see them made obsolete. This unusual convergence of interests has made itself apparent repeatedly during discussions about the potential uptake of basic income in European countries, he says.
Lehto believes the most important area of focus for academics who wish to see basic income implemented should be on the development of more concrete models. He also wants to see research on how guaranteed income might find its way into national constitutional frameworks, so that it won’t end up the victim of short-term political interference.
Universal Basic Income and the Future of Human Endeavour
While artificial intelligence poses a threat to some jobs in its current form, it looks like most of us will still have a place in the workforce for the foreseeable future. So, is that likely to change, and if so, how soon?
The crucial question here is whether there are aspects of human ingenuity that AI simply won’t be able to replicate. If there are, then human workers may indeed have a path to continued relevance in the long run, even if our functions are drastically different than at present.
This is the question of artificial general intelligence, a construct Wikipedia defines as an “intelligent agent… that could learn to accomplish any intellectual task that human beings or animals can perform.”
According to deep learning and AI expert Kelvin Lwin, this type of all-powerful tool is still a while away. He believes that the exponential growth in AI that some commentators have predicted is unlikely to materialize in the immediate future. “We’re probably at the peak in terms of these models… they’re running out of data, essentially.” Lwin thinks currently available tools will improve in terms of “specific tasks,” but that we’re unlikely to continue seeing forward leaps like those that have happened over the last year.
Another major issue is energy. Lwin highlights that the power consumption of current artificial intelligence capabilities is massive; given that the world is already in the depths of an energy crisis, this is an intractable problem.
Lwin compares AI’s likely current trajectory with projections around the dotcom boom in the early 2000s. He highlights the fact that, regardless of the heady predictions people made about the potential of the internet at the time, it still took us almost two decades to get services like Amazon’s next-day delivery.
Lwin does believe that businesses will do their best to replace human labor with software, as this is where the greatest potential value of the technology lies. However, he is skeptical about how easy this will be in the immediate term.
He points to driving work as an example, noting that previous projections predicted fully automated vehicles by 2020. In reality, we are still many years from seeing driverless cars take over the road in 2023.
Ultimately, Lwin says, AI will eventually start to replace more and more jobs and progress to a point at which it emulates every type of human intelligence. It just won’t happen as soon as many think.
The Way Forward
So, the marriage of universal basic income and artificial intelligence may not be as straightforward as first imagined. The uptake of automation may indeed spur new economic productivity that doesn’t rely on human labor, and this should make it easier for governments to redistribute wealth. Right now, though, the roadblocks (both technological and political) in the way of basic income may be too large.
As Otto Lehto points out, it’s just a really big change. While “small reforms” are relatively easy, dramatic shifts require “a lot of cooperation between different parties [and] institutions.” This is far from common in Western bureaucracies. The obstacles are not just financial; they’re also related to old habits that are hard to kill. The popularity of basic income as a concept, and the growing body of evidence suggesting its effectiveness, may change this, but it won’t happen overnight.
So, universal basic income may indeed arrive, and automation might help it along, but we’re not on the cusp of some idle utopia facilitated by robots who attend to our every need. Human work and traditional welfare systems are here to stay, for the time being at least.