While it's tempting to think that leading economists, senior academics and best-selling authors are not prone to sweeping elevator pitches, Simon Johnson – who is all three – is keen to describe his latest book as "the thousand-year backstory for ChatGPT". Johnson is one half of the authorship duo (his co-author is the equally well-credentialled Daron Acemoglu) that has just delivered 'Power and Progress'. He says that the "slightly longer" version of the one-line summary relates to how forms of automation and machines that have been with us for more than a thousand years have been used "to replace labour and make humans essentially more productive".
The subtitle of 'Power and Progress' reveals how Johnson and Acemoglu's analysis of automation has historically brought with it a "struggle over technology and prosperity". Although this is to simplify the scope of the authors' work, its central theme is that while technology has always been supposed to make everyone's lives better – that it is a "tide that lifts all boats" – what we've seen for much of the millennium is that a tiny elite routinely becomes more empowered and distanced from the workers, who experience little of the anticipated prosperity that was promised.
In the Middle Ages, says the former IMF chief economist, advances in agricultural techniques funded centuries of cathedral building while the peasantry (90 per cent of the population) remained on the breadline. During the Industrial Revolution, a minority of capitalists amassed the wealth of Croesus while workers toiled long hours in harsh conditions for subsistence wages that barely increased over centuries. We have a version of this in the AI-empowered 21st century, in which the rank and file at FAANG-style organisations have yoghurt fridges and ping-pong tables, while their bizarrely out-of-touch billionaire bosses indulge in fantasies of commercial space travel for their celebrity chums.
Such milestones are common enough, although there are times, such as the late 19th to the first three-quarters of the 20th century, "when things went much better, when real wages did rise and living standards increased. Not for everyone, but for most people. That was a sharing of prosperity that was a remarkable achievement and the culmination of a long struggle".
Image credit: Dreamstime
Johnson is alluding to an era exemplified by automobile manufacture when the 'bandwagon of progress' (see extract below) "seemed to work. But then we lost it. After 1980 we did not maintain that pattern of shared prosperity." Which means that "the question we are grappling with today is what happens now? Can technology be redirected? Is there a way to redirect generative AI and other innovations in a way that allows us to better share what comes next?"
There is an existing template for those wishing to 'change visions' about how to redirect technology to create shared prosperity. If we accept, as the authors argue, that "our current problems are rooted in the enormous economic and social power of corporations, especially in the tech sector" – whose "most pernicious" impacts are felt in automation, surveillance, data collection and advertising – then we could do a lot worse than to reflect on what happened in turn-of-the-20th-century America.
"To regain shared prosperity we must redirect technology, and this means activating a version of the same approach that worked more than a century ago for the Progressives." This was an era of 'muckraker' investigative journalism in magazines such as McClure's that campaigned for social reform. It was also an era that contributed to structurally altering the narrative and norms. And we need this today, says Johnson: "Society and its powerful gatekeepers need to stop being mesmerised by tech billionaires and their agenda. Debates on new technology ought to centre not just on the brilliance of new products and algorithms but on whether they are working for the people or against the people."
Power and Progress
To economists Daron Acemoglu and Simon Johnson, if the past millennium has taught us anything it is this: progress is not automatic. It depends, they say in their superb 'Power and Progress', on the choices we make about technology. This is because new ways of harnessing innovations often fail to fulfil their boasts of becoming the bedrock of widespread prosperity; instead, frequently serving the narrow interests of an elite. It's always been the case. Early industrialisation in Britain created vast fortunes for the few, while workers' wages stagnated. Today, digital technologies such as AI and the internet 'increase inequality and undermine democracy'. But it doesn't have to be this way, the authors argue. Technology can be bought under control, but only if we take major decisions out of the hands of a technocracy that exists mostly to elevate its own power, wealth and prestige.
Redirection – "changing people's minds and providing alternative visions" – is one of the key concepts of 'Power and Progress', and Johnson reckons that the case of the Progressives is "one of the most informative", especially when it came to breaking up giant corporations – something he thinks should happen today. And yet muckraker journalism (a century ago the term was nowhere near as pejorative as it is today) is remembered more for its tactic of sensationalism than its effectiveness: "People say that Ida Tarbell either exaggerated or didn't get her facts straight, and so on... but who cares? She changed people's minds about trusts and Big Oil and John D Rockefeller. And, of course, the really interesting – and very American – thing about breaking up Standard Oil was that the shareholders made money, Rockefeller became richer and we got more competition. There are some win-wins to be had."
There are recent examples of redirection in action. One of Johnson's favourites is the "focus and effectiveness of AIDs campaigners at the end of the 1980s. They were arguing for more research and development to address HIV/AIDS: that was their main ask. And I remember thinking at the time: 'why is that what they want?' In retrospect I understood that they wanted to redirect those federal science dollars to focus on solving the problem," with the result that treatments were found.
Johnson stresses that redirection need not be simply about bringing down capitalists to create an alternative system. "It's about more transparency and treating consumers more fairly. Today, in the context of generative AI, let's think about who controls the 'means of information'. Karl Marx went on about the 'means of production' and he had some good points, although societies didn't work out how he expected. What we're looking at now is how Microsoft and Google have a duopoly on the means of information and those foundation Large Language Models."
'Power and Progress' posits that without breaking up such colossal tech giants, "things could play out badly for us, particularly because they've got these blinkers on. The computer science field and tech executives are totally obsessed with machine intelligence, which is double-speak for 'let's fire some more people'."
'Power and Progress: Our Thousand Year Struggle Over Technology and Prosperity' by Daron Acemoglu and Simon Johnson, Basic Books, £25
The bandwagon of progress
The 'productivity bandwagon' ... maintains that new machines and production methods that increase productivity will also produce higher wages. As technology progresses, the bandwagon will pull along everybody, not just entrepreneurs and owners of capital.
Economists have long recognised that demand for all tasks, and thus for different types of workers, does not necessarily grow at the same rate, so inequality may increase because of innovation. Nevertheless, improving technology is viewed as the tide lifting all boats because everyone is expected to derive some benefits. According to the conventional wisdom, to rectify the rise in inequality and build even more solid foundations for shared prosperity, workers must find a way to acquire more of the skills they need to work alongside new technologies.
The theory behind the 'productivity bandwagon' is straightforward: when businesses become more productive, they expand their output. For this, they need more workers, so they get busy with hiring. When many firms attempt to do so at the same time, they bid up wages. For example, in the first half of the 20th century one of the most dynamic sectors of the US economy was car manufacturing. As Ford Motor Company and then General Motors (GM) introduced new electrical machinery, built more efficient factories and launched better models, their productivity soared, as did their employment. From a few thousand workers in 1899, producing just 2,500 automobiles, the industry's employment to rose than more than 400,000 by the 1920s. By 1929, Ford and GM were each selling around 1.5 million cars every year.
Edited extract from 'Power and Progress' by Daron Acemoglu and Simon Johnson, reproduced with permission.
