But could AI and robotics companies do the R&D and sell automation technologies to other companies? To the extent that companies constrain their operations to meet robots’ limitations, yes. But even still, human labor is necessary to configure, calibrate, and adjust automation technologies to adapt to a changing world, whether those changes are a differently shaped product or a bird that flies into the factory. Brynjolfsson and McAfee lump together a wide swath of AI applications and predict that the successes among them portend the more general expansion of automated work. But in doing so they overlook the enormous amounts of behind-the-scenes, domain-specific labor that makes AI possible in the first place. Google’s self-driving car doesn’t simply go anywhere its passengers please. For this car to drive “itself,” a human worker has to drive around, scan, and map the car’s world — including everything from curb heights to intersection angles. Machine-learning algorithms that partially automate data processing still need to be trained for every new form, or every new kind of topic the algorithm might deal with. Other robots profiled in The Second Machine Age will learn the movements of shop floor workers and then replace them, until the next tune-up or calibration is necessary. Such work of alignment is not a bug — it is the condition of possibility for keeping humans and automation working in the same world. The Second Machine Age leans heavily on the accounts of corporate executives promising fantastic new horizons of tech profit, but it’s undeniable that for those pursuing customers and venture capital for automation, there’s good money to be had in hiding these headaches.
This care for and feeding of artificial intelligence suggests a much bigger oversight in Brynjolfsson and McAfee’s argument. Automation doesn’t replace labor. It displaces it. Historian Ruth Schwartz Cowan famously showed how the invention of the washing machine mainly increased the standards of cleanliness domestic workers (paid and unpaid) had to meet. Shoshana Zuboff’s 1988 book In the Age of the Smart Machine described how factory automation created text-based labor, displacing workers who smelled and felt wood pulp with those who could read screens and meters and tend to the machines. Hamid Ekbia and Bonnie Nardi call these managerially advantageous human-machine configurations “heteromation.”2
The emergence of the digital microwork industry to tend artificial intelligence shows how labor displacement generates new kinds of work. As technology enterprises attempt to expand the scope of culture they mediate, they have had to grapple with new kinds of language, images, sounds, and sensor data. These are the kinds of data that flood Facebook, YouTube, and mobile phones — data that digital microworkers are then called on to process and classify. Such microworkers might support algorithms by generating “training data” to teach algorithms to pattern-match like a human in a certain domain. They might also simply process large volumes of cultural data to prepare it to be processed in other ways. These cultural data workers sit at computer terminals, transcribing small audio clips, putting unstructured text into structured database fields, and “content moderating” dick pics and beheadings out of your Facebook feed and Google advertisements.3
Computers do not wield the cultural fluencies necessary to interpret this kind of material; but people do.4 This is the hidden labor that enables companies like Google to develop products around AI, machine learning, and big data. The New York Times calls this “janitor work,” labeling it the hurdle, rather than the enabling condition, of our big data futures.5 The second machine age doesn’t like to admit it needs help.