The Economist breaking positions to sound the alarm on AI

Tecologists tend to predict that the economic impacts of their creations will be unprecedented – and this is especially true when it comes to artificial intelligence. Last year, Elon Musk predicted that continued advances in AI would make human work obsolete. OpenAI CEO Sam Altman written that AI will inevitably continue the shift of economic power from labor to capital and create “phenomenal wealth.” Jensen Huang, CEO of semiconductor design company Nvidia, compared Development and deployment of AI for a “new industrial revolution”.

But while technologists are optimistic about the economic impacts of AI, members of that other technocratic priesthood with profound influence on public life – economists – are not. Even if technologists create the powerful AI systems they say they will soon, the economic impacts of those systems will likely be daunting, many economists say. According to for Tyler Cowen, professor of economics at George Mason University, AI “will increase the US annual growth rate by between a quarter and a half percentage point”. Nobel Prize winner in Economics, Paul Krugman, argued that “history suggests that the big economic effects of AI will take longer to materialize than many people currently seem to expect.” David Autor, professor of economics at the Massachusetts Institute of Technology, written that “the industrialized world is flooded with jobs and will remain that way”.

As a professor of economics at the University of Virginia, affiliated with some of the profession’s most highly regarded institutions, Anton Korinek is a certified member of the priesthood. The 46-year-old, however, broke the rankings. He uses his discipline’s methods to model what would happen if AI developed as many technologists say it could—meaning that AI systems will likely outperform humans at any task by the end of the decade. In the sterile lexicon of academic economics, he paints an alarming picture of a near future in which humans will no longer play a role in the economy and inequality will increase.

If “the complexity of the tasks that humans can perform is limited and full automation is achieved, then wages will collapse,” write Korinek and one of his Ph.D. students, Donghyun Suh, in a recent paper. In other words, if there is a limit to the complexity of tasks that humans can perform, and machines become advanced enough to fully automate all of these tasks, then wages could fall drastically as human labor is no longer necessary. , leaving anyone who doesn’t share in the resources that go back to AI systems and their owners to starve.

Economists have spent the last 200 years explaining to the uninitiated that the labor fallacy — the idea that there is a fixed number of jobs and that automating any one of those jobs will create permanent unemployment — is wrong, says Korinek. “In a way, this is our professional background,” he says. “We spend so much time fighting a false narrative that it’s hard to change when the facts actually change and see that this situation can in fact be different.”

“The scariest part is that the technological predictions I use as data are those that people like Sam Altman or [Anthropic CEO] Dario Amodei preaches freely in public everywhere,” says Korinek. “I am not a technology expert and have no special insight or knowledge on these issues. I’m just taking what they say seriously and asking: What would this really mean for the economy, for jobs and for wages? It would be completely disturbing.”

Korinek’s father, a primary care physician, instilled in him a passion for understanding the brain. As a teenager in Austria, Korinek became fascinated by neuroscience and learned to program. In college, he took a course on neural networks—the brain-inspired AI systems that have since become the most popular and powerful in the industry. “It was intriguing, but at the same time boring, because the technology just wasn’t very powerful using the tools we had at the time,” he recalls.

When Korinek was finishing college in the late 1990s, the Asian financial crisis hit. The 1997 crisis, precipitated by a currency crisis in Thailand, followed many other episodes of acute financial difficulties in emerging markets. Korinek decided to pursue a doctorate. about financial crises and how to avoid them. Fortunately for him – though less fortunately for everyone else – the global financial crisis hit in 2008, a year after he completed his thesis. His experience was demanded and his career took off.

As an academic economist, he watched from the sidelines as AI researchers made breakthrough after breakthrough; scanning newspapers, listening to radio programs and reading books about AI. In the early 2010s, as AI systems began to outperform humans in tasks such as image recognition, he began to believe that artificial general intelligence (AGI) – the term used to describe an AI system – still to be built, which could perform any task that a human being could perform. could – could be developed alarmingly soon. But it took the birth of his first child, in 2015, to take him away from the scenes. “There’s really something happening, and it’s happening faster than I thought in the 1990s,” he remembers thinking. “This will really be relevant to my daughter’s life trajectory.”

See more information: 4 Charts That Show Why AI Progress Is Unlikely to Slow Down

Using the methods of economics, Korinek sought to understand how future AI systems could affect economic growth, wages and employment, how inequality created by AI could endanger democracy, and how policymakers should respond to the economic questions posed. by AI. For years, working in the AGI economy was a marginal activity. Economics Ph.D. students told Korinek they would like to work on the issues he was beginning to study, but felt they needed to stick to more conventional topics if they wanted to get a job. Despite gaining tenure in 2018 and therefore being insulated from the pressures of the academic job market, Korinek sometimes felt discouraged. “Researchers are still social creatures,” he says. “If you only face skepticism about your work, it becomes much, much harder to push a research agenda, because you start to second-guess yourself.”

Until OpenAI launched the wildly popular chatbot, ChatGPT, in November 2022, Korinek says research into the economics of AGI was a fringe activity. Now, it’s “on an exponential growth path,” says Korinek.

Many economists still rule out, quite rightly, the possibility of AGI being developed in the coming decades. Skeptics point to the many ways in which current systems fall short of human capabilities, the potential obstacles to AI’s continued progress, such as a scarcity of data to train larger models, and the history of technologists making overconfident predictions. A paper published in February by economists in Portugal and Germany dismisses AGI as “science fiction” and therefore argues that AI is unlikely to cause explosive economic growth.

See more information: When will AI be able to surpass us? Depends on who you ask

“Two years ago, this was absolutely common – this was the average reaction,” says Korinek. But more people are starting to consider the possibility of developing AGI, and those who continue to dismiss the possibility are providing solid arguments for their position, rather than rejecting it out of hand. “In other words, people are getting very seriously involved in the debate,” he says approvingly.

Other economists are open to the possibility of AGI being developed in the near future, but argue that it would not yet precipitate a collapse in employment. They often attribute disagreement with those who think the development of AGI could cause these things to economic illiteracy on the other side.

For example, in March, economist Noah Smith argued in a blog post which was captured by new York Times that high-paying jobs will be plentiful even after AI can outperform humans at any task. This happens, Smith explains sparingly, because AI systems and humans can specialize in tasks where their relative advantage is greatest – an economic idea known as comparative advantage – and then trade with each other. When “most people hear the term “comparative advantage” for the first time, they immediately think of the wrong thing,” Smith writes.

But Korinek is hardly economically illiterate. “I say I understand comparative advantages,” he says with a smile. Korinkek’s models, which consider comparative advantage, still do not paint an optimistic picture for humans in a world where AI systems can perform any task a human can do. “The theory of comparative advantage tells us that there is room for gain from trade,” says Korinek. “But the big problem is that it doesn’t tell us that the terms of that trade will be sufficient for humans to make a decent living, or even for humans to cover their subsistence costs.”

Later in his blog post, Smith goes on to say that government intervention, in the form of a tax on AI or setting limits on the amount of energy AI systems can use, could be necessary to ensure that humans can gain enough to survive. if AI can outperform them in any task. But while Smith is confident that governments will take these steps if necessary, Korinek is less certain.

“Economists are slowly starting to take the potential job disruption caused by AI more seriously,” he says. “Policymakers aren’t there yet – that’s my concern. I’m nervous about it because we might not have a lot of time.”

Korinek does not share the certainty with which many technologists predict the imminent development of AGI. But given the potential implications of such technology, society must prepare, he says.

“It could be many decades before we come close to replicating the things the human brain does,” he says. “But I think there’s also a high enough chance that it really won’t take that long and that it will happen very quickly, and that it will turn our society upside down.”

More from TIME