After Ludditism

In September 2017, during a moment of tense hyperpartisanship unrivaled in the United States since the end of the Civil War, the House of Representatives nonetheless unanimously approved a measure promoting the sale of autonomous vehicles by exempting them from certain federal regulations and safety standards. At the same time, they forbade any of the fifty states from banning such vehicles. The fact of this unanimity is, by itself, a remarkable registration of the techno-optimism that still undergirds American ideology, even in the pessimism of the so-called Anthropocene; once a proposition has been marked as "the future," which is always imagined as a future of luxury, wealth accumulation, and consumer convenience, it becomes incredibly hard in the U.S. to even begin to articulate any possible position of resistance or opposition to it.

That the House bill has since stalled in the Senate is also a remarkable turn of events: concerns over the diminishment of safety standards and the eroding of consumers' privacy and redress rights, as well as a fundamental ambiguity about what it means to require that a self-driving car be equivalent in safety to a human-piloted vehicle, have left the bill languishing in a Senate committee through the end of the 115th Congress. (Undoubtedly a pair of high-profile fatalities involving self-driving cars on public roadways in March 2018—one from an Uber test vehicle and the other from Tesla—also played a role in this stalemate, as did the chaotic government shutdown that prevented any last-minute resurrection of the bill in December 2018.) The possibility that the emergence of a future we have been told is inevitable and irresistible can be interrupted, even momentarily, comes as something of a shock. Still, whatever the precise regulatory framework in the United States ultimately turns out to be, there seems little doubt that self-driving vehicles remain imminent. Audi is already selling a vehicle in Germany with a "Traffic Jam Pilot" that can essentially drive itself on a divided highway at speeds over 35 miles per hour, with drivers able to take their hands off the wheel and their attention from the road. The company will implement this feature in their U.S. fleet as soon as it is legal to do so.

Self-driving cars present only one special case of the move toward automation and algorithmic intelligence that is now transforming leisure, labor, policy, and policing across the globe. Natural language processing networks can already do work in law, finance, human resources, advertising, content production, and other industries that were once the purview of trained human experts. Machine-learning algorithms monitor our emails, cursor movements, shopping patterns, and social interactions in the name of stitching together a total profile of our behavior to serve us slightly better targeted digital ads. Facial recognition technology, microexpression analysis, and ubiquitous surveillance by security forces in public spaces borders on what Philip K. Dick once called "Precrime."¹ In education, there is the promise of university lectures without professors and student papers that grade themselves. Signs of this impending autonomous future are everywhere we look.

Grim acceptance of the supplanting of the human by artificial intelligences of one sort of another, and anticipation of a future of abandonment and deprivation for those left behind, seems to be the only possible response, even among the hype-men and buzz-marketers who welcome the change. Take, for instance, the bitter prognostication of economist Tyler Cowen in his widely circulated 2013 Politico essay "The Robots Are Here," speaking as a proponent of the coming digital meritocracy:

The rise of intelligent machines will spawn new ideologies along with the new economy it is creating. Think of it as a kind of digital social Darwinism, with clear winners and losers: Those with the talent and skills to work seamlessly with technology and compete in the global marketplace are increasingly rewarded, while those whose jobs can just as easily be done by foreigners, robots or a few thousand lines of code suffer accordingly.²

Nor does this transformation promise an easy reduction to existing class biases about what we consider to be "skilled...