Only yesterday, artificial intelligence was still the stuff of science fiction; now, it casts a portentous shadow over the future of work. Depending upon which breathless commentator one believes, AI promises to relieve us of the tedious aspects of our work — or threatens to deprive us of our jobs entirely. Seeking historical perspective, I reached for the classic account of the evolution of the labor process under capitalism, Harry Braverman’s 1974 Labor and Monopoly Capital.

Braverman’s book ranges further, and sees more deeply, than its blunt subtitle, “The Degradation of Work in the Twentieth Century,” might suggest. Like his acknowledged model, Marx’s description of the transformation of the production process in Capital, Braverman provides a meticulous investigation of the restless making and remaking of the organization of labor under capitalism. But he never loses sight of the impact of these serial upheavals on the working class.

Braverman rejected simplistic interpretations of Marx as a technological determinist. Rather, he points out that a new invention always presents an array of possibilities. In the near term, the dominant social relations shape which of these possibilities are cultivated and which actively foreclosed. Capitalist relations of production exhibit an “incessant drive to enlarge and perfect machinery on the one hand, and to diminish the worker on the other.” This dynamic reflects capitalism’s larger tendency to separate conception from execution — the work of the brain and the work of the hand. The result is a small stratum of highly trained (and handsomely paid) professionals on the one side and a swelling mass of proletarianized laborers condemned to mindless tasks on the other.

Braverman brought a singular perspective to his investigation. He had apprenticed as a coppersmith and subsequently found employment in the steel industry, earning his living as a craftsman for fourteen years before cofounding a newspaper, the American Socialist. (He spent the remainder of his career in publishing, directing the storied independent socialist imprint the Monthly Review Press until his death in 1976.) Despite the rapid decline of the coppersmithing trade in which he was trained, Braverman bristled at the inference that his criticisms reflected nostalgia for an antiquated past: “Rather, my views about work are governed by nostalgia for an age that has not yet come into being.” Braverman’s background in the trades, as well as his decades-long involvement in socialist activism, made him uniquely equipped to take the baton from Marx and extend Capital’s analysis of the labor process into the twentieth century.

The pivotal figure in Labor and Monopoly Capitalism’s narrative is Frederick Winslow Taylor (1856–1915), the eccentric founder of the scientific management movement. From his childhood, Taylor showed signs of extreme obsessive-compulsive disorder, counting his steps and seeking ever-more efficient ways to perform the most mundane activities. “These traits fitted him perfectly for his role as the prophet of modern capitalist management,” Braverman deadpans, “since that which is neurotic in the individual is, in capitalism, normal and socially desirable for the functioning of society.”

So long as workers directed the labor process, Taylor maintained, they would never perform “a fair day’s work” — which he defined, naturally, as the maximum amount that they could perform without injury. Therefore, capitalists must not rest content with owning the means of production and the commodities labor produced: they needed to control the labor process itself. Taylor tends to be remembered for squeezing greater productivity out of workers by prescribing their every movement in accordance with the dictates of his “science.” But, Braverman suggests, his more important feat was to systematically compile the craft knowledge that had hitherto belonged to labor and transfer it to management.

Soon, workers were left performing simplified detail work that had been decontextualized from the production process as a whole; meanwhile, management enjoyed a monopoly on the technical know-how that, historically, had been the patrimony of the skilled trades. The ongoing separation of the conception and execution of labor that characterizes production under capitalism had reached a new threshold. This process subsequently repeated itself in management, creating a handful of corner-office executives and an army of deskilled administrative assistants and middle managers.

Labor and Monopoly Capital tells a sobering story, but by no means an unhopeful one. Braverman detected signs of capitalism’s historical limits in the fact that new technology frequently reunites and automizes the steps of the labor process that the division of labor had fragmented. In his final lecture, delivered in the spring 1975, Braverman urged that “Workers can now become masters of the technology of their process on an engineering level and can apportion among themselves in an equitable way the various tasks connected with this form of production that has become so effortless and automatic.” Liberated from the drudgery of repetitive tasks thanks to automation, a team of associated producers might reclaim the unity of the production process once enjoyed by craftworkers on a higher plane.

AI holds out a similar possibility of reuniting, in automated form, many of the skills and bodies of knowledge that the capitalist division of labor has pulverized in its relentless quest for control and efficiency. If predictions that AI will inaugurate an age of universal leisure are wildly optimistic, the prospect that socialized workers might direct the entirety of the production process with its assistance seems less so.

But we will have to fight for it. Capitalism customarily takes advantage of technological advances by firing workers and demanding greater productivity from the few it does not cull. Braverman informs us that the verb “to manage” “originally meant to train a horse in his paces, to cause him to do the exercises of the manège.” Management has always viewed the labor process as a site of struggle, and it is determined to keep hold of the reigns. If we want AI to improve rather than replace or further degrade our jobs, a reading of Braverman suggests that we must be prepared to carry the battle into the very labor process itself.