One Economist’s Case for Taking AI Seriously

When Anton Korinek sat down for this interview, he let his AI agent make the introduction. “My research focuses on how AI is reshaping labor markets, inequality, and economic growth, and what that means for policy,” the chatbot said. Korinek, a professor of economics at UVA’s College and Graduate School of Arts and Sciences and Darden School of Business, had considerably more to add.

Korinek spent his early career studying financial crises, but he had a longstanding interest in neural networks. In 2015, watching AI improve rapidly while also becoming a new father, he decided
to shift his focus to the economic implications of AI. Before he joined UVA, he was upfront with his future colleagues about the pivot. “Just to be sure that you are aware, I am shifting my research towards AI,” he told them. They were supportive.

At UVA, he connected with the Human and Machine Intelligence Group, an interdisciplinary community of researchers studying AI across fields. “It has helped me keep pace with what’s going on in AI and find an intellectual community that is interested in these questions beyond my core field of economics,” he said.

His focus sharpened around artificial general intelligence (AGI), the still-theoretical point at which AI reaches human-level capability across the board. Despite rapid recent advances, AI has not yet significantly moved the economic needle. “From big macro perspectives, you can’t really see AI anywhere yet,” Korinek said. Even so, investment in AI has grown by a factor of 2.5 every year for the past two years. “By some measures, AI production has been growing by more than 2,000% per year,” he said. “The staggering sums being invested in server farms are hard to justify if AI remained where it was last year.”

Korinek’s research centers on two questions: how to distribute the economic benefits of AGI broadly, and how to improve the economic models that will help us understand what AI transformation actually looks like.

His latest paper asks what happens when AI research itself becomes automated.

The headline result is that if we fully automate AI research, we may see a singularity, a really significant growth explosion within as little as three years.

AKAnton KorinekProfessor of economics at UVA’s College and Graduate School of Arts and Sciences and Darden School of Business

“The headline result is that if we fully automate AI research, we may see a singularity, a really significant growth explosion within as little as three years,” Korinek said.

Korinek has been named to the TIME100 AI list of the most influential people in artificial intelligence and is currently heading research on transformative AI at Anthropic. At UVA, he serves as Faculty Co-Director of the Economics of Transformative AI Initiative (EconTAI), which he describes as being targeted at “not yesterday’s AI, but tomorrow’s AI.”

EconTAI also organizes scenario planning workshops across UVA, pushing faculty and students in every field to reckon with what AI transformation means for their own work. “My hope is to inspire as many as possible among them to use their subject matter expertise to think about how rapid advances of AI will affect their field,” Korinek said.

This story originally appeared in the 2026 Annual Research Report, produced by the Office of the Vice President of Research. Find more stories from this report here: https://news.research.virginia.edu/category/annual-report/

Video Transcript:

 Artificial intelligence is not only a technology, but it is something that will fundamentally transform our world, including our economy. If you look at the macroeconomic statistics, growth numbers, productivity, unemployment, you cannot see an impact of artificial intelligence quite yet, and that has been puzzling because in our personal interactions, we can sense that AI has already become quite powerful.

Companies are still figuring out how to optimally employ these tools, and in the cognitive sector, it is difficult to measure productivity. The idea of artificial general intelligence that can broadly perform all the cognitive functions that our human brains can perform, and we could have a rapid intelligence explosion if the AI

is well-aligned with us humans, then it could usher in a world that is, you know, close to utopian. On the other hand, it could lead to, I think, disastrous outcomes on a scale that is also difficult to imagine. In the coming few years, artificial intelligence will completely change our world.