A friend sent me a Politico editorial by Tyler Cowen about the way intelligent machinery will further erode the middle class, and otherwise upend the world economy. Some additional searching revealed that this editorial and an interview with the American Enterprise Institute are really instruments for publicizing his 2013 book Average Is Over: Powering America Beyond the Age of the Great Stagnation. I'd be interested to read the entire monograph, but these two pieces at least give me a rough idea of what he's interested in. The gist seems to be that our economic future will be split between the minority (Cowen keeps calling it the 15%) who can design and collaborate with intelligent machinery, and the hoi polloi forever resigned to working for or tending to the 15%'s automated apparatuses.
The rise of the machines builds on deeper economic trends that are already roiling American society, including stagnant growth since 2001 and a greater openness to trade and foreign outsourcing. But it’s the rapid increase in machines’ ability to substitute for intelligent human labor that presages the greater disruption. We’re on the verge of having computer systems that understand the entirety of human “natural language,” a problem that was considered a very tough one only a few years ago. We’re close to the point when we can fit the (articulable) knowledge of the entire world into the palm of our hands. Self-driving cars are making their way onto streets in California and Nevada. Whether you are a factory worker or an accountant, a waitress or a doctor, this is the wave that will lift you or dump you.
The warning seems primarily directed at so-called thinking professions that have generally enjoyed increasing job numbers and wages during the 20th century—a trend only recently changed. Perhaps the hope is that highlighting the threat this economic shift poses to people with money and influence (i.e. to white collar, intellectual workers) will bring about some intervention in or reform of the labor market—intervention that was rather lacking as industrial robots and outsourcing unemployed factory workers were unemployed en masse during the 20th century.
What strikes me most about the editorial and interview is how little it really says about the future (the speculation regarding the capabilities of yet-to-invented computer systems is no more or less believable than the Technological Singularity). The two pieces use the robot-bogey man to bring our attention back to existing problems, so I'm tempted to read them more like sci-fi romans à clef. We already see the problems Cowen points to, for example, in San Francisco. Indeed, it is the city he seems to have in mind during his interview with the American Enterprise Institute, because SF nicely demonstrates the divide between those people creating and selling automated information processing machinery—along with the service industry of tutors, designers, restauranteurs, etc. catering to them—and the remainder, the passed over—the working class, the bohemians, etc. NPR even made the Bay Area the focus of a special series on growing income inequality. Presumably as the intelligent machine migrate out of Silicon Valley and into the wider world, they bring with it the Bay Area's new social caste system.
Cowen, however, is hardly the first to warn about automation causing unemployment and social unrest. The mathematician and cyberneticist Norbert Wiener made similar predictions when he published Cybernetics in 1948—automated machinery threatened to displace factory workers and thereby render thousands unemployed with scant prospects. That's precisely what happened in the United States. Unlike Cowen, Wiener was also concerned the machines might eventually grow beyond human control and enslave us.1 A softer version of the robot takeover appears in some of Asimov's robot stories (Cowen opens with a quotation from Asimov), in which humans simply allow more benevolent and efficient intelligent machines to take over the drudgery of governing and running the world economy.
Cowen's prediction of these intelligent machines driving income inequality has a literary precedent as well. The introduction of mass-produced humanoid robots generates a new class divide in Karel Čapek's 1920 drama R.U.R. (1920), which gave us the word robot. The scientist industrialists who designed the robots and own the company manufacturing them hold all the world's capital, because their product is the only one worth buying. As the robots replace human workers, the profoundly unemployed masses cause social unrest, and so the world's governments purchase robot militaries to quell the revolts. As with most Faustian and Sorcerer's Apprentice tales, the creatures turn on their masters and eliminate humanity, since they find us to be inefficient, lazy, and otherwise a drain on the planet's limited resources.
Wiener, unlike Cowen, was optimistic that these automated laborers could, if used properly, liberate humans from manual toil and thereby democratize the leisure time needed for creativity and innovation.2 It is this liberation from responsibility and freedom to think, create, and otherwise build that, I suspect, fosters the libertarian tendencies of many technology entrepreneurs and enthusiasts in Silicon Valley. Once we strip away the rhetoric and speculation about intelligent machines replacing physicians and lawyers aside, we see that Cowen actually wants to ensure this libertarian tendency does not operate as an excuse for one class of people to horde the new freedom automation brings. Wiener assumed such equality, but Cowen is skeptical of that optimism, more in the tradition of R.U.R.
I would, however, disagree with the strong implication in the two pieces that intelligent machines entail a libertarian ideology. The two are strongly connected because many tech-company owners espouse some form of libertarian thinking, and because proponents of the Technology Singularity that Cowen correctly associates with machine thinking replacing human thinking has a libertarian flavor. Two key Singularity theorists and boosters have libertarian tendencies, even if they do not explicitly identify with any libertarian political groups. Vernor Vinge's fiction depicts the future as a self-regulating economy that needs little or not government intervention; and Ray Kurzweil's discussion of technology enhancing individual freedom appeals to the libertarian belief in autonomous individuals freed from limitations (though in Kurzweil's case the limits are not political so much as biological).3 The connection between intelligent machinery, the information economy, etc and libertarian thinking is may simply be a historical accident, rather than a logical necessarity.
The same technology that produces machines equal or greater in intellect than any human would allow not just for a libertarian paradise of free markets, individual autonomy and other radical freedoms, but a surveillance and control system beyond even what we see today. In such a world, like the one the Asimov story referenced earlier, human decisions and action would only appear free given our limited perspective and intellect—more capable intelligences would manipulate the economy.
Indeed, the very notion that intelligent machines would replace doctors, lawyers, college professors, and even the software engineers themselves undermines the very concept of individuality that Cowen would need for his libertarian dystopia. Who needs humans at all when machines can everything we do but more efficiently? Vinge, at least, admits that these super-intelligent machines could just as easily decide to do away with humanity—no matter how much money some humans may have accumulated.4 Besides, if we're talking about the future, we should also make room for the possibility that other events may intervene to prevent the economic dystopia Cowen envisions, such as meteor strike, Socialist extraterrestrials, the rapture, etc.
Instead of wringing our hands about the libertarian dystopia heralded by intelligent machinery, we should instead consider ways of rehabilitating Wiener's vision of liberating human potential through such automation. Frithjof Bergmann's New Work comes close to doing that without explicitly mentioning Wiener. Bergmann isn't advocating an abolition of capitalism and free markets, so much as shifting the emphasis away from unsustainable growth and ever increasing efficiency. It is indeed the drive for efficiency that means eliminating the lest efficient part of most workflows—humans. Instead, Bergmann imagines we could use new technologies (like 3D printing) to provide a basic standard of living for ourselves, work fewer hours (because we will need less money to get by), and use the newly freed time to pursue personal or community projects. "The Partially Examined Life," a new favorite podcast of mine, devoted two episodes to Bergmann's work and interviewed him.
- A recent Wiener biography—Norbert Wiener: Dark Hero of the Information Age—notes that he tried to bring the danger to the attention of labor unions. ↩
- See The Human Use of Human Beings (1950). ↩
Vinge's "Realtime" series depicts humans just before the Technological Singularity as occupying a highly, self-regulating, laissez-faire economy. The works include The Peace War (1984), "The Ungoverned" (1985), and Marooned in Realtime (1986). They were collected as Across Realtime (2000).
- Reference to his piece on the sinularity. ↩