Getting the regulation of artificial intelligence right is one of the most urgent problems facing our species, and also one of the most delicate. AI has the potential to improve most aspects of our lives; Alphabet chief executive officer Sundar Pichai argues that its impact will be “more profound than electricity or fire.”
It also has the potential to damage us profoundly; in one survey of AI researchers, 48 percent thought that there was at least a 10 percent chance that its impact would be “extremely bad,” that is, lead to human extinction.
How can we maximize the upside of the new technology and minimize the downside?
This is the subject at the heart of an important new book by two prominent economists. Daron Acemoglu and Simon Johnson are professors at the Massachusetts Institute of Technology, one of the leading temples in the cult of tech. Acemoglu is the co-author (with James Robinson) of “Why Nations Fail: The Origins of Power, Prosperity and Poverty.” Johnson is a former chief economist at the International Monetary Fund. In “Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity,” they look at a millennium of technological innovation to understand the likely impact of AI.
The answer they come to is not cheerful, though they’ve reached that conclusion by way of an irritating Brahmin populism. It’s a book written for people who hang out in anointed innovation districts like Kendall Square in Cambridge, Mass., or in the privileged halls of Harvard.
Acemoglu and Johnson dismiss the techno-optimistic view that technology inevitably brings progress in its wake at the heart of a certain type of liberalism. They say there is nothing automatic about new technologies bringing widespread prosperity.
Throughout history, powerful elites have seized control of new technologies and used them to enrich themselves and extend control over their subordinates. This is not just a matter of extracting the surplus generated by improvements in productivity. It involves skewing the ways that technology is developed and applied to benefit one group over another.
Here are some examples from the book. Improvements in agriculture during the Middle Ages–better plows, crop rotation and mills–enriched landlords and clergy while often leaving the peasants worse off. Eli Whitney’s cotton gin, which significantly improved the productivity of the cotton industry by making it easier to separate the plant’s fiber from its sticky green seeds, helped to entrench slavery and extend its adoption in the U.S. The tech revolution from the 1980s onward has made bosses rich while keeping workers’ income flat, thanks to a combination of outsourcing, re-engineering and ideology.
The authors concede that technological progress is often the work of challengers to the status quo. The British industrial revolution was driven by “the middling sort” of self-educated artisans who revolutionized production with steam while the elites were peacocking around. George Stephenson, inventor of the Rocket train, was the son of poor illiterate parents in Northumberland. Richard Arkwright, whose innovations revolutionized the textile industry, was a tailor’s son. But these technologies and their makers were eventually co-opted by the ruling class.
Countervailing forces can come along and redirect technology from elite enrichment to creating shared gains. The authors praise the combination of electoral competition, trade union power and reforming intellectuals and politicians. Yet the authors worry that AI is exploding on to a world where such forces have been emasculated. Business titans enjoy more power and prestige than they have since the gilded age, organized labor is puny, and democracy has been captured by money. The winning formula (innovation plus guidance) has been replaced by a losing one (let the elites control technology).
In Acemoglu and Johnson’s view, the digital revolution has already been hijacked by self-seeking elites. The computer hackers’ dream world of distributed power and open innovation has been replaced by a hellscape of oligopoly of tech giants. These use machines and algorithms to replace workers: they monitor employees to squeeze more surplus value out of them.
“One of the things we hear consistently from workers is that they are treated like robots in effect because they’re monitored and supervised by these automated systems,” said one labor advocate cited in the book.
The new oligopoly has created surveillance capitalism: an economic system that gathers ever more information on all of us to sell to advertisers. These advertisers, along with media moguls, can use this information to manipulate the masses more effectively than they have ever been manipulated before, personalizing ads, shaping the information environment, and playing with people’s emotions. The result is a fundamental challenge to John Stuart Mill’s 19th-century notion of the sovereign individual.
The authors’ main worry about AI is not that it will do something unexpected like blowing up the world, though that would be undesirable. It is that it will supercharge the current regime of surveillance, labor substitution and emotional manipulation. Their grand solution is to use public policy to refocus the new technology from “machine intelligence” to “machine usefulness.” But they warn that before we can have a chance of doing that, we need to educate public opinion and recharge democracy.
The book proposes an interesting set of policies to produce a better version of the future: Provide government subsidies to develop more socially beneficial technologies; refuse to give patents to technologies aimed at worker or citizen surveillance; eliminate tax incentives to replace labor with machines; break up big tech companies that enjoy market shares not seen since the days of the American industrialists John D. Rockefeller and Andrew Carnegie; repeal Sector 230 of the 1996 Communications Decency Act that protects Internet platforms against legal action or regulation because of the content they host, and impose a digital advertising tax.
The belief that something must be done about tech is not as original as Acemoglu and Johnson seem to think. The very first sentence of the book is: “Every day we hear from executives, journalists, politicians that we are heading relentlessly toward a better world, thanks to unprecedented advances in technology.” Actually, every day most of us hear the opposite.
There is a lot of popular and media anxiety over AI. Elon Musk has joined dozens of tech luminaries calling for a six-month pause in the creation of the most advanced forms while we come to terms with its implications. Henry Kissinger, the master strategist who recently turned 100, worries that AI, being an efficiency-maximizer, will steer future military conflicts in the direction of unparalleled savagery.
Rather than the one-way trip to high-tech perdition conjured up by the authors, we may be in the middle of making an important choice; indeed, reaffirming Mill in the end.
There is a kind of willful myopia to the book. Acemoglu and Johnson say little about the benefits that tech innovation brings to consumers. The most striking thing about 19th-century innovations, for example, is not their impact on wages, as the authors assert, but their impact on the quality of life in general. People who lived in the dark could summon up light after sunset, thanks to electricity. People who had never been more than a few miles from their homes could travel across the country, thanks to the railway.
Technology’s role as a liberator was accelerated in the 20th century: Think of the role of the radio in bringing entertainment to isolated farms, or the washing machine and vacuum cleaner in reducing the amount of time spent on domestic drudgery. These benefits are not the result of benevolent intellectuals dividing up the surplus for the common good, but of capitalists pursuing profit by selling people what they want.
The authors have a distracting way of dividing the world into elites (bad) and the people (good). In truth, the elites include plenty of reformers: free traders like Robert Peel, whose repeal of the Corn Laws in 1846 ushered in the age of affordable breakfasts. The people are not always angels. Trade unions have been obstacles to the introduction of new technologies. The British printing unions, which fought for years to prevent the introduction of electronic printing, were notorious for restrictive practices and employing ghost workers.
Acemoglu and Johnson don’t recognize the extent to which “the people” can sometimes act as vested interests rather than apostles of the common good. The difficult trick is to strike a balance between allowing the market to produce its (often unforeseen) benefits through competition and preventing it from being distorted by special interests. This can only be done by having a clear-eyed view of both “the people” and “the elites.”
For all its sins, Big Tech has provided us with electronic wonders that put much of the world’s knowledge at our fingertips. AI is already beginning to do the opposite of what the authors say the oligopoly is plotting: It is empowering regular workers by making it easier to find and present information, providing us all with our own research assistant.
The authors rightly worry about the way the Chinese government is using the digital revolution to monitor and repress its people. But what about India? Thanks largely to Nandan Nilekani, a tech billionaire and chairman of Infosys, India has introduced the world’s biggest biometric ID system that provides 1.3 billion Indians with a digital identity. People who once had no way of proving their identity now have access to unemployment payments, bank accounts and mobile services. This has simultaneously revolutionized the lives of the poor and increased the state’s ability to monitor the population.
The history cited in the book as part of its argument is also crude. The assertion that “England produced little of lasting value during the entire medieval period” might surprise admirers of Oxford and Cambridge universities, Chaucer’s writings, or Magna Carta.
It’s a pity the book is so rough-hewn. That’s because Acemoglu and Johnson highlight a substantial worry about the evolution of the tech industry. The emancipatory power of the Internet has undoubtedly been reduced by the marriage between Google and advertising (something that founders Larry Page and Sergey Brin never contemplated as students). The Internet is now used as much to tempt us to buy stuff we don’t need as it is to democratize information. The emancipatory power of AI will surely be limited and distorted in the same way.
But the authors’ perspective limits the appeal of the book. In their view, “hate speech” always spews forth from “white nationalists,” never from, say, the anarchists and antifa activists who turned Portland into a riot zone. There are many people on the right who are equally worried about the power and direction of tech. Conservatives are also concerned about the ability of tech companies to get rich by plugging directly into the basest sides of our nature. The best way to produce a new regulatory regime is to build a broad coalition that includes the right.
There is nothing inevitable about the direction of technology. Powerful people can direct it toward narrow interests rather than the common good. Clearsighted coalitions of the concerned can lead it in more enlightened ways. Time may be running short given the pace of AI’s advance but there is still time to save ourselves from digital slavery.