The future of work is increasingly on our minds. Now that AI is coming for our jobs, will we end up supervising or being supervised by it? One way of spending the time freed up by smart tech is to read Control Science, an economic history showing how work rules were established and have since come to dominate our lives.

The book’s timeline covers the past 400 years, its settings ranging across the world from North America to Europe to Japan and back to the US. A historian of labour, Henry Snow dissects four entrenched ideas: that society is a mere collection of individuals; that they are solely driven by selfishness; that they are therefore incapable of self-administered planning; and that ‘everything is – and should be – a market’. Technocrats have long argued that these tenets underpin the world, but as this study demonstrates, the reality is far more complex.

One of the earliest case studies included here concerns the Bentham brothers’ experiments in the late 18th century. While Samuel championed exploitation through innovation, Jeremy pondered ways to ‘produce happiness through profit’. In the 1790s, the Benthams came up with the Panopticon, a circular structure in which inspectors could watch inmates without making it obvious. Although never fully implemented in an industrial environment, this ‘project of its day, conceived [by Samuel] in absolutist Russia and expanded in counter-revolutionary Britain’, became a symbol of prison. The Panopticon is still in use today on a scale likely to astonish even Jeremy’s auto-icon at UCL, in the form of digital surveillance.

The chapter entitled ‘Political Economy in Antebellum America, 1800-65’ introduces ‘a broader transatlantic conversation about how the few could control the many’. Early on, ‘political economists claimed that the social world was as predictable and controllable as the natural world’. This argument was undermined when black southerners ended slavery by going on what was, to quote W.E.B. Du Bois, a ‘general strike’. ‘The plantation was thus undone by the kernel of truth at the heart of classical political economy: all wealth does come from labour.’

American economic relations form a prequel to early exchanges between Japan and the West, the subject of the next chapter. After Japan, isolated for centuries, was opened by American gunboats in 1854, some of its efficiency advocates were influenced by the Victorian philosopher Herbert Spencer, ‘the most influential 19th-century proponent of the science of control’. Other Japanese thinkers, however, disagreed with his tendency to compare economic forces with natural ones. What they were facing, after all, was ‘a man-made calamity, one engineered and profited from’. Looking for solutions, a Japanese mission visited London in 1872. The scenes its members witnessed in the East End made one of them remark: ‘The whole world appears despicable.’ Consequently, Japan went down a different path, relying less on laissez-faire competition, an approach that allowed it to fight off ‘colonial ambition but not capitalism’.

In order to remain in power, 19th-century capitalists ‘would need their science of control to pass as a real science’, writes Snow. When I was studying management and applied mathematics in the 1990s, many still wondered whether management qualified as a science. No such doubts arose about another subject, artificial intelligence. Back then it was unfashionable to think of AI as something that could become yet another power mechanism. ‘The shift from tools that aided workers to machines that workers aided’, which Snow traces to three centuries ago, is now fully manifest in our relationship with technology. At a recent conference, I asked a speaker about AI’s potential to help human researchers and was told that the question has already been reversed: a computer linguist’s research, to give one example, is valued mostly for its ability to be used to train large language models.

After excursions into the 20th century, with its Taylorism and price theory, macroeconomics and decline of trade unions, the final chapter takes Snow to our times, and to Amazon. He visits the company’s fulfilment centres – contemporary versions of the Panopticon – noting that, although there are worse employers around (among large retailers, Walmart is particularly notorious), this ‘precisely engineered system of surveillance and optimisation’ is ‘an exploitative mess’. The overview of Jeff Bezos’s empire is followed by brief encounters with other tech bros. ‘Beff Jezos’ stands out as a particularly pushy peddler of control science ‘wrapped in self-important techno- babble’. Ultimately it is this pseudoscientific construct that ‘paved Donald Trump’s road to the presidency… Why not elect an utterly selfish man without virtue? He knew how the world really works. Art of the deal!’

In the 19th century, ‘managers wanted long hours and machine-like consistency from workers’. Today, many of them still do. While practical aspects of work are constantly changing, the conflict between the managers and the workers – the controlling and the controlled – remains. It may seem eternal; yet Snow’s analysis helps to show us (to quote the title of his newsletter) another way.