Power and Progress

The Co-Directors of the MIT Shaping the Future of Work Initiative, Daron Acemoglu, David Autor, and Simon Johnson, have released a plan to choose the best path for AI development, answering the question: “Can We Have Pro-Worker AI?” Read the policy memo.

From the summary, here are the five most important federal policies to pursue at this time:

1. Equalize tax rates on employing workers and on owning equipment/algorithms to level the playing field between people and machines.
2. Update Occupational Safety and Health Administration rules to create safeguards (i.e., limitations) on the surveillance of workers. Finding ways to elevate worker voice on the direction of development could also be helpful.
3. Increase funding for human-complementary technology research, recognizing that this is not currently a private sector priority.
4. Create an AI center of expertise within the government, to help share knowledge among regulators and other officials.
5. Use that federal expertise to advise on whether purported human-complementary technology is appropriate to adopt in publicly provided education and healthcare programs, including at the state and local level.

Research that underlies this policy memo has been supported by many different funding organizations over the last decade. Acemoglu gratefully acknowledges financial support for related projects from Accenture, the Air Force Office of Scientific Research, the Army Research Office, the Bradley Foundation, the Canadian Institute for Advanced Research, the Department of Economics at MIT, Google, the Hewlett Foundation, IBM, Microsoft, the National Science Foundation, Schmidt Sciences, the Sloan Foundation, and the Toulouse Network on Information Technology. Autor gratefully acknowledges financial support for related projects from Google, the Hewlett Foundation, the NOMIS Foundation, and the Washington Center for Equitable Growth. Johnson gratefully acknowledges support from the Sloan School, MIT.

Prologue
What Is Progress?

Every day, we hear from executives, journalists, politicians, and even some of our colleagues at MIT that we are heading relentlessly toward a better world, thanks to unprecedented advances in technology. Here is your new phone. There goes the latest electric car. Welcome to the next generation of social media. And soon, perhaps, scientific advances could solve cancer, global warming, and even poverty.

Of course, problems remain, including inequality, pollution, and extremism around the globe. But these are the birth pains of a better world. In any case, we are told, the forces of technology are inexorable. We couldn’t stop them if we wanted to, and it would be highly inadvisable to try. It is better to change ourselves—for example, by investing in skills that will be valued in the future. If there are continuing problems, talented entrepreneurs and scientists will invent solutions—more-capable robots, human-level artificial intelligence, and whatever other breakthroughs are required. People understand that not everything promised by Bill Gates, Elon Musk, or even Steve Jobs will likely come to pass. But, as a world, we have become infused by their techno-optimism. Everyone everywhere should innovate as much as they can, figure out what works, and iron out the rough edges later.

WE HAVE BEEN here before, many times. One vivid example began in 1791, when Jeremy Bentham proposed the panopticon, a prison design. In a circular building and with the right lighting, Bentham argued, centrally positioned guards could create the impression of watching everyone all the time, without themselves being observed supposedly a very efficient (low-cost) way of ensuring good behavior.

The idea at first found some traction with the British government, but sufficient funding was not forthcoming, and the original version was never built. Nevertheless, the panopticon captured the modern imagination. For the French philosopher Michel Foucault, it is a symbol of oppressive surveillance at the heart of industrial societies. In George Orwell’s 1984, it operates as the omnipresent means of social control. In the Marvel movie Guardians of the Galaxy, it proves to be a flawed design that facilitates an ingenious prison breakout.

Before the panopticon was proposed as a prison, it was a factory. The idea originated with Samuel Bentham, Jeremy’s brother and an expert naval engineer then working for Prince Grigory Potemkin in Russia. Samuel’s idea was to enable a few supervisors to watch over as many workers as possible. Jeremy’s contribution was to extend that principle to many kinds of organizations. As he explained to a friend, “You will be surprised when you come to see the efficacy which this simple and seemingly obvious contrivance promises to be to the business of schools, manufactories, Prisons, and even Hospitals. . . .”

The panopticon’s appeal is easy to understand—if you are in charge—and was not missed by contemporaries. Better surveillance would lead to more compliant behavior, and it was easy to imagine how this could be in the broader interest of society. Jeremy Bentham was a philanthropist, animated by schemes to improve social efficiency and help everyone to greater happiness, at least as he saw it. Bentham is credited today as the founder of the philosophy of utilitarianism, which means maximizing the combined welfare of all people in society. If some people could be squeezed a little in return for a few people gaining a great deal, that was an improvement worth considering.

The panopticon was not just about efficiency or the common good, however. Surveillance in factories implied inducing workers to labor harder, and without the need to pay them higher wages to motivate greater effort.

The factory system spread rapidly in the second half of the eighteenth century across Britain. Even though they did not rush to install panopticons, many employers organized work in line with Bentham’s general approach. Textile manufacturers took over activities previously performed by skilled weavers and divided them up more finely, with key elements now done by new machines. Factory owners employed unskilled workers, including women and small children, to perform simple repetitive tasks, such as pulling a handle, for as many as fourteen hours per day.They also supervised this labor force closely, lest anyone slow down production. And they paid low wages.

Workers complained about conditions and the backbreaking effort. Most egregious to many were the rules they had to follow in factories. One weaver put it this way in 1834: “No man would like to work in a power-loom, they do not like it, there is such a clattering and noise it would almost make some men mad; and next, he would have to be subject to a discipline that a hand-loom weaver can never submit to.”

New machinery turned workers into mere cogs. As another weaver testified before a parliamentary committee in April 1835, “I am determined for my part, that if they will invent machines to supersede manual labour, they must find iron boys to mind them.”

To Jeremy Bentham, it was self-evident that technology improvements enabled better-functioning schools, factories, prisons, and hospitals, and this was beneficial for everyone. With his flowery language, formal dress, and funny hat, Bentham would cut an odd figure in modern Silicon Valley, but his thinking is remarkably fashionable. New technologies, according to this view of the world, expand human capabilities and, when applied throughout the economy, greatly increase efficiency and productivity. Then, the logic goes, society will sooner or later find a way of sharing these gains, generating benefits for pretty much everybody.

Adam Smith, the eighteenth-century founding father of modern economics, could also join the board of a venture capital fund or write for Forbes. In his view, better machines would lead to higher wages, almost automatically:

In consequence of better machinery, of greater dexterity, and of a more proper division and distribution of work, all of which are the natural effects of improvement, a much smaller quantity of labour becomes requisite for executing any particular piece of work, and though, in consequence of the flourishing circumstances of the society, the real price of labour should rise very considerably. . . .

In any case, resistance is futile. Edmund Burke, contemporary of Bentham and Smith, referred to the laws of commerce as “the laws of nature, and consequently the laws of God.”

How can you resist the laws of God? How can you resist the unstoppable march of technology? And anyway, why resist these advances?

ALL OF THIS optimism notwithstanding, the last thousand years of history are filled with instances of new inventions that brought nothing like shared prosperity:

• A whole series of technological improvements in medieval and early modern agriculture, including better plows, smarter crop rotation, more use of horses, and much improved mills, created almost no benefits for peasants, who constituted close to 90 percent of the population.

• Advances in European ship design from the late Middle Ages enabled transoceanic trade and created massive fortunes for some Europeans. But the same kinds of ships also transported millions of enslaved people from Africa to the New World and made it possible to build systems of oppression that lasted for generations and created awful legacies persisting today.

• Textile factories of the early British industrial revolution generated great wealth for a few but did not raise worker incomes for almost a hundred years. On the contrary, as the textile workers themselves keenly understood, work hours lengthened and conditions were horrible, both in the factory and in crowded cities.

• The cotton gin was a revolutionary innovation, greatly raising the productivity of cotton cultivation and turning the United States into the largest cotton exporter in the world. The same invention intensified the savagery of slavery as cotton plantations expanded across the American South.

• At the end of the nineteenth century, German chemist Fritz Haber developed artificial fertilizers that boosted agricultural yields. Subsequently, Haber and other scientists used the same ideas to design chemical weapons that killed and maimed hundreds of thousands on World War I battlefields.

• As we discuss in the second half of this book, spectacular advances in computers have enriched a small group of entrepreneurs and business tycoons over the last several decades, whereas most Americans without a college education have been left behind, and many have even seen their real incomes decline.

Some readers may object at this point: Did we not in the end hugely benefit from industrialization? Aren’t we more prosperous than earlier generations, who toiled for a pittance and often died hungry, thanks to improvements in how we produce goods and services?

Yes, we are greatly better off than our ancestors. Even the poor in Western societies enjoy much higher living standards today than three centuries ago, and we live much healthier, longer lives, with comforts that those alive a few hundred years ago could not have even imagined. And, of course, scientific and technological progress is a vital part of that story and will have to be the bedrock of any future process of shared gains. But the broad-based prosperity of the past was not the result of any automatic, guaranteed gains of technological progress. Rather, shared prosperity emerged because, and only when, the direction of technological advances and society’s approach to dividing the gains were pushed away from arrangements that primarily served a narrow elite. We are beneficiaries of progress, mainly because our predecessors made that progress work for more people. As the eighteenth-century writer and radical John Thelwall recognized, when workers congregated in factories and cities, it became easier for them to rally around common interests and make demands for more equitable participation in the gains from economic growth:

The fact is, that monopoly, and the hideous accumulation of capital in a few hands, like all diseases not absolutely mortal, carry, in their own enormity, the seeds of cure. Man is, by his very nature, social and communicative—proud to display the little knowledge he possesses, and eager, as opportunity presents, to encrease his store. Whatever presses men together, therefore, though it may generate some vices, is favourable to the diffusion of knowledge, and ultimately promotive of human liberty. Hence every large workshop and manufactory is a sort of political society, which no act of parliament can silence, and no magistrate disperse.

Electoral competition, the rise of trade unions, and legislation to protect workers’ rights changed how production was organized and wages were set in nineteenth-century Britain. Combined with the arrival of a new wave of innovation from the United States, they also forged a new direction of technology—focused on increasing worker productivity rather than just substituting machinery for the tasks they used to perform or inventing new ways of monitoring them. Over the next century, this technology spread throughout Western Europe and then the world. Most people around the globe today are better off than our ancestors because citizens and workers in early industrial societies organized, challenged elite-dominated choices about technology and work conditions, and forced ways of sharing the gains from technical improvements more equitably.

Today we need to do the same again.

The good news is that incredible tools are available to us, including magnetic resonance imaging (MRI), mRNA vaccines, industrial robots, the internet, tremendous computational power, and massive amounts of data on things we could not measure before. We can use these innovations to solve real problems—but only if these awesome capabilities are focused on helping people. This is not the direction in which we are currently heading, however.

Despite what history teaches us, the predominant narrative today has shifted back toward something remarkably close to what was prevalent in Britain 250 years ago. We are living in an age that is even more blindly optimistic and more elitist about technology than the times of Jeremy Bentham, Adam Smith, and Edmund Burke. As we document in Chapter 1, people making the big decisions are once again deaf to the suffering created in the name of progress.

We wrote this book to show that progress is never automatic. Today’s “progress” is again enriching a small group of entrepreneurs and investors, whereas most people are disempowered and benefit little.

A new, more inclusive vision of technology can emerge only if the basis of social power changes. This requires, as in the nineteenth century, the rise of counterarguments and organizations that can stand up to the conventional wisdom. Confronting the prevailing vision and wresting the direction of technology away from the control of a narrow elite may even be more difficult today than it was in nineteenth-century Britain and America. But it is no less essential.

Impacts of AI, Regulatory Frameworks

The MIT Shaping the Future of Work Initiative has calculated preliminary estimates of the workforce effects from automation driven by generative AI platforms like ChatGPT. Read the policy memo. Here is an excerpt:

Roughly 1.6–3.2 million workers could lose their jobs over the next 20+ years, around 1–2% of total US employment via May 2022 OEWS. These are gross job losses, we expect many of these people will find new employment. Based on the observed effects from other, similar employment shifts in US history, this may coincide with wage declines of ~33–47% for the demographic groups which are engaged in occupations most-exposed to generative AI, relative to others who are less exposed to this form of automation. 

Initiative Co-Director Daron Acemoglu has a new NBER working paper, cowritten with Todd Lensman, proposing a framework for “Regulating Transformative Technologies.” Here is a summary:

Transformative technologies like generative AI promise to accelerate productivity growth across many sectors, but they also present new risks of potential misuse. We develop a multi-sector technology adoption model to study the optimal regulation of transformative technologies when society can learn about these risks over time. Socially optimal adoption is gradual and convex. If social damages are proportional to the productivity gains from the new technology, a higher growth rate leads to slower optimal adoption. Equilibrium adoption is inefficient when firms do not internalize all social damages. Sector-independent regulation is helpful but generally not sufficient to restore optimality. 

View Presentation Slides

View the Power and Progress slide deck that Daron Acemoglu and Simon Johnson have used to present their findings to a range of audiences.

Sample the Audiobook

Hear the official excerpt from HachetteAudio on SoundCloud, as read by Malcolm Hillgartner.

Listen to the full audiobook on your platform of choice.

Read an Excerpt

“The antidemocratic turn of technology, and how we turn it back.” 

Available via ChangeThis from Porchlight Books.

“It’s time to rechart the course of technology. Here are 4 ways to start.”

Available via Ideas Made to Matter from MIT Sloan.