On August 11, MIT Nobel laureate economist Daron Acemoglu published his latest book, What Happened to Liberal Democracy?: Remaking a Politics of Shared Prosperity.
For much of the second half of the 20th century, liberal democracy appeared to be remarkably successful. Economic growth raised living standards across broad segments of society, technological advances created new industries and good jobs, governments expanded public services, and democratic institutions gave citizens both political rights and a meaningful voice in public affairs.
But, something has clearly gone wrong.
Across many advanced democracies, economic inequality has increased, working-class communities have struggled, political polarization has intensified, and confidence in democratic institutions has declined. Why did a political and economic system that once enjoyed such broad support lose the confidence of so many of its citizens?
“The crisis of liberal democracy is the defining challenge of our age. Our lives would be transformed for the worse without democracy and the basic freedoms that liberal democratic institutions enshrine,” said Acemoglu.
“The irony is that liberal democracy was once very popular because it delivered what people wanted: shared prosperity, true voice in public affairs, and high quality public services. This crisis is rooted in liberal democracy, in theory and in practice, turning its back on these founding aspirations.”
Acemoglu argues that “we need a new governing philosophy for democracy” and proposes working-class liberalism as the foundation of that philosophy. It would prioritize shared prosperity and issues that matter to the whole population rather than primarily to a narrow elite. Liberal democracy renewed in this way, he added, would provide the appropriate framework for addressing challenges ranging from AI and climate change to shifting global balances.
The Breakdown of Shared Prosperity
In a recent article referencing Acemoglu’s book, The Economist noted that for many decades liberal democracy met voters’ expectations by spreading the benefits of innovation and economic growth through higher wages, taxes, and public services.
But over the past few decades, that relationship began to weaken. Computers and the Internet contributed to growing inequality by “raising the demand for brains over brawn.” Automation, offshoring, globalization, and related economic changes reduced the availability of good jobs for many workers without college degrees.
At the same time, according to The Economist, increasingly influential college-educated liberal elites embraced cultural values that alienated significant numbers of working-class voters, contributing to a political backlash and increasingly polarized and toxic political debates.
A few days after his book’s publication, Acemoglu explored the growing divide between college-educated and working-class Americans in “How College Education Divided America,” an article published in Time.
He recalled flying to Washington, D.C., a few years ago for a meeting with left-leaning foundations on a topic close to his heart: combating inequality and reversing widening earnings disparities, particularly “the declining wages of workers, and especially men, without college degrees.”
“To my surprise, nobody seemed to comment on the stagnation and decline in the wages of non-college workers,” he wrote. “Nor was there any discussion of how inequality between residents of small cities and large metropolitan areas had surged.”
“Instead, commentator after commentator talked about gender inequality, racial inequality, and inequality affecting sexual minorities. These are all important topics, of course, but if our aim was to understand the surge in inequality at the aggregate level, much of the story was about jobs and opportunities for Americans without college education.”
Acemoglu isn’t arguing that inequalities involving race, gender, or sexual orientation are unimportant. Rather, his point is that the growing economic and geographic divide between Americans with and without college degrees has not received comparable attention, despite its importance in explaining broader economic inequality and political discontent.
“The political rise of the college educated is related to the phenomenal growth in the number of people with college degrees,” he added. “But even more important, it is a consequence of the rising status of the college educated, which is itself rooted in their meteoric economic climb — while automation, offshoring, globalization, and related sectoral trends have reduced the availability of good jobs for non-college workers.”
The divide isn’t only economic.
“As the college educated have become politically more powerful, they have also diverged further from the less educated in their geographic locations and their social lives, reducing their ability to empathize with the latter’s problems and creating a major cultural rift between the two groups,” wrote Acemoglu.
“In the process, a coalition between different segments of the college-educated and the tech sector has formed. I would argue this coalition has completely turned its back on the economic priorities of the less educated and has come to focus on cultural and global issues.”
AI: A New Technological Crossroads
Artificial intelligence could now significantly amplify these trends — or potentially help reverse them.
In principle, AI could make virtually everyone richer by supercharging innovation and productivity. But AI could also amplify some of the negative consequences associated with earlier digital technologies by displacing large numbers of workers, further concentrating income and wealth, and contributing to what The Economist described as “a full-blown crisis of democracy.”
This is where Acemoglu’s analysis of liberal democracy connects with his extensive research on technology and work.
AI isn’t simply another economic issue confronting democracies. Depending on how it’s developed and deployed, it could either deepen the economic forces that have contributed to today’s political divisions or help create a new era of broadly shared prosperity.
In his new book, Acemoglu argues for a “pro-worker AI” — finding ways to use AI to make workers more productive and their labor more valuable.
For example, AI could enable electricians to troubleshoot increasingly sophisticated electrical grids, using sensors to spot problems and databases of previous glitches and solutions to help humans analyze them. Or AI could help teachers by assessing students’ answers to quizzes and suggesting how to divide classes into smaller groups for more individualized instruction, thereby complementing rather than replacing human teachers.
The objective isn’t to protect every existing job or prevent automation. Rather, it is to direct more technological innovation toward creating new tasks and capabilities in which human labor becomes more productive and valuable.
Acemoglu nicely explained what he means by pro-worker AI in his keynote on “The Long Term Evolution of AI Economics” at the 2025 MIT Sloan CIO Symposium, which I wrote about in this previous blog post.
Automation versus Human Complementarity
He opened his keynote by noting that we face two very important choices in the deployment of AI, with major economic and cultural implications for the future.
The first is automation technologies — AI tools aimed primarily at automating and replacing human labor in order to lower labor costs. Automation can reduce the prices of products and services, but “the bottom line for workers would not be that great.”
The second is human-complementary technologies — AI tools designed to enhance workers’ capabilities and productivity, enabling them to perform more sophisticated tasks that they might not otherwise have been able to do.
“The future will have both sorts of technologies,” said Acemoglu. The balance between the two will create different kinds of winners and losers and “is going to determine the broader impacts of AI.”
From a purely economic perspective, the bias toward automation is understandable. Managers under pressure to reduce costs in highly competitive markets are naturally attracted to technologies that offer quick and relatively predictable returns.
“There’s nothing wrong with automation,” said Acemoglu. “If we had not automated some of the manual tasks that people used to perform in the 18th century, we would live in a very different world.”
But companies that make history generally do a lot more than reduce costs.
Acemoglu cited the Ford Motor Company, which helped revolutionize manufacturing by taking advantage of the growing deployment of electricity in the early 20th century. Electrification wasn’t transformative simply because factories replaced steam engines with electric motors. Its greater impact came when companies redesigned factories and production processes around electricity’s capabilities, including assembly lines composed of decentralized electric machines and supported by increasingly skilled workers.
“But the question is whether there are now enough incentives in the tech sector to actually develop the human complementary machines,” said Acemoglu, because they may be less profitable and less attractive in the short term than simpler automation technologies.
“I think many managers will be very willing to find ways of using technology that are going to make their most valuable human resource, skilled employees, more productive. Even some of them may be imaginative enough to find ways of making their less skilled workers more productive.”
But ultimately, he argued, “the economics creates a small or moderate bias towards too much automation and not enough of these human complementary technologies that can create these new services, new tasks and new products.”
Adding to the bias toward automation are what Acemoglu has called the “fever dreams” surrounding artificial general intelligence, or AGI. While the definition of AGI remains somewhat unclear, the aspiration to develop AI systems capable of performing essentially all the cognitive tasks humans can perform naturally points technological development toward replacing rather than complementing human labor.
Technology, Shared Prosperity, and Democracy
This technological choice has major implications that extend well beyond productivity and employment.
If AI further concentrates income and opportunity among highly educated workers and owners of capital while reducing the economic prospects of everyone else, it could deepen precisely the economic, geographic, and cultural divisions that Acemoglu argues have weakened liberal democracy.
But if AI helps create better jobs and makes a much broader range of workers more productive, technological progress could once again become an important contributor to shared prosperity.
This is why his argument about the direction of AI is ultimately about more than technology.
“In the age of AI, will we build tools that displace people or ones that expand human potential? Will we chase narrow efficiency or pursue shared prosperity?” asked Acemoglu near the end of his CIO Symposium keynote.
“I think one of the very important things about the future of AI is to learn from the past of digital technologies, but also dream big about how companies can revolutionize themselves in terms of offering new services, new products, new tasks, and going into history books.”
The connection between technological change and liberal democracy may be one of the most important messages of Acemoglu’s new book.
Liberal democracy cannot survive on political rights and elections alone. Its long-term legitimacy also depends on whether large numbers of citizens believe that the economic and political system works for them — that technological progress creates opportunities in which they can participate, that their work is valued, that their children can reasonably expect better lives, and that they have a meaningful voice in the decisions affecting their future.
AI could make that challenge considerably worse. If its primary economic impact is to automate human work, concentrate wealth, and further increase the divide between highly educated professionals and everyone else, it could reinforce some of the forces that have contributed to today’s crisis of liberal democracy.
But AI can also be developed as a human-complementary technology — helping workers acquire expertise, perform more sophisticated tasks, create new products and services, and become more productive and valuable.
The great challenge of the emerging AI era may therefore be about much more than building increasingly intelligent machines. It may be about using those machines to help rebuild the shared prosperity on which successful liberal democracies have historically depended.