Automation and the Future of Work - Aaron Benavav, Verso Books (2020) & The AI Economy - Roger Bootle, John Murray Press (2019)
These books review a singular feature of modern economies – workers are competing with ever more advanced machines. Both authors also question the idea that technological unemployment can be addressed by a Universal Basic Income (UBI).
The advent of Chat GPT sparked a media frenzy around two inter-related questions – does artificial intelligence (AI) represent a qualitative break from the past, and, if so, will the future of work be human-less?
Looking at the issue from contrasting perspectives, Benavan and Bootle (both of whom focus their analysis on wealthy countries) reach similar conclusions – they doubt the hype. Starting with Benavav, he sees the puzzle of slowing jobs growth as a precedent to AI.
Automation theorists contend that AI-driven productivity gains have dislodged millions of jobs. Benanav offers a longer-term, neo-Marxist view, pointing out that global productivity has been falling since the 1970s, well before the advent of Chat GPT.
He suggests that the slowdown in global output per worker has resulted from an overcrowded playing field – as industrial competition between countries escalated after the 1950s, lower prices undercut profits and disincentivized investment.
Factories in advanced economies shuttered and manufacturing jobs with decent wages became harder to find. On the other hand, part-time, low-wage and insecure work became more normalised.
Compared to low-skilled service jobs (like hospitality and retail), menial manufacturing (factory floor work) benefits from higher productivity due to economies of scale. Wages also benefit from higher basic training and greater union representation.
Benavan argues that techno-optimists have thus confused correlation for causation. Falling demand for labour is not the result of AI, but a function of long-term economic stagnation. In turn, he sees post-manufacturing capitalism as moribund and crisis-stricken.
In slow-moving economies, he does concede that technological improvements can lead to job destruction. In particular, he anticipates more manufacturing unemployment as a result of automation (factory tasks can be more easily optimised than service sector work).
More broadly, he views the economic realities of the Global South – pervasive underemployment paired with a large low-wage service sector and significant inequality – as the future awaiting Western countries (with “our country first” nationalism a political by-product).
Moreover, he doesn’t see UBI as a solution to “late capitalism.” Though UBI would provide the ‘have-nots’ with just enough to live above the poverty line, it would also allow owners of capital to continue enriching themselves without limit, facilitating even more inequality.
Still, Benavav fails to convincingly pivot from his unsparing diagnosis to a compelling account of political action. He conforms to type, advocating for a socialistic future based on centralised economic planning and the dissolution of private property.
Roger Bootle (arguably the City of London’s best-known economist) examines the problem in reverse. Rather than query Silicon Valley’s grasp of economic history, he disputes its vision of the future.
From the outset, he appears determined to dismiss the possibility of AI-induced unemployment, arguing instead that that robots will (largely) complement human labor, not substitute it.
As with the emergence of China in the 1990s, Bootle admits that AI will produce a degree of structural unemployment. Overlaps with Benavav’s thesis end there, as Bootle goes on to note that any contractionary impact from AI will be short-term and “transitional”.
Over the long-term, Bootle argues that AI will raise average wages. His view rests on two assumptions: first, many jobs stand to benefit from AI-linked productivity gains through increased computing power; second, jobs destroyed by robots will be replaced by new ones.
For winners in the AI economy (data scientists and robot manufacturers), many will wish to reduce their working hours. Bootle envisages a growing demand for leisure-linked activities and attendant jobs in areas like entertainment, travel and fitness coaching.
Healthcare work will also proliferate, as the relative size of elderly populations continues to grow. More generally, humans will maintain an absolute advantage over robots in tasks requiring manual dexterity and emotional intelligence (or non-linear reasoning).
Bootle’s sweeping book, ranging in scope from AI’s impact on bond yields to critiques of Thomas Piketty’s work on inequality, tries to steer a course between techno-utopia and dystopia. For all his insights, though, he dismisses the risks to low-paid work too easily.
He states that “if there is greater inequality [from automation], it will not be on a significant scale.” In turn, he appears to welcome the expansion of an insecure labor force and ignores the socio-political impacts from worker precarity.
A passing familiarity with America’s rustbelt offers a sobering warning to the notion that free markets can smoothly manage structural unemployment.
Predictably, for someone who views AI as labor augmenting, Bootle’s counter-arguments to UBI are unconvincing – it would require a “debilitating” increase in taxes, and would offend “most people’s sense of fairness.”
That said, he doesn’t take a complete Panglossian view. He stresses that governments should reform public education systems in light of increased roboticization through greater funding for digital training and the provision of “lifetime learning”.
Bootle and Benavav foresee a future with more insecure work. They also agree that UBI is misguided (though neither make a compelling case). The market cheerleader sees it as an unnecessary cost to a manageable problem that price signals can correct, while the Marxist doubts it will solve a system in crisis.
And so, everything has to change for things to stay the same.
