SILICON VALLEY — An economist who studies the economic impact of artificial intelligence at Google DeepMind warned that economies falling behind in AI will struggle to achieve even 2% annual growth, citing Europe as an example. His point was that AI is the reason the U.S. continues to post growth in the 2% range even as low growth becomes entrenched worldwide.
Alex Imas, director of AGI economics at DeepMind, said 2% growth would be a difficult target to hit without AI innovation, speaking on the 14th at an event unveiling research findings at the company's headquarters in Mountain View, California. Imas previously worked as a professor of behavioral science and economics at the University of Chicago Booth School of Business and now leads the economics research team at DeepMind.
In a conversation with James Manyika, Google's senior vice president of research, technology and society, Imas said: "The expectation that we'll grow 2% every year is treated almost as a stylized fact, based on 150 to 200 years of well-organized GDP data — but that's not guaranteed." He added: "We have a prime example of this. Europe, which was on the same growth path as us through the 1990s and early 2000s, has now completely fallen off that path."
"If you look closely at that 2% growth rate, it contains an enormous amount of essential innovation, science and technology, and productivity-enhancing elements," Imas said. "If there are obstacles, or if innovation slows, 2% growth does not actually materialize." He continued: "When people ask me why the economy isn't showing the kind of growth AI is seeing, with 13% or 14% expansion, I tell them we are in fact growing 2%. That is a big win. It's because things are much lower on the other side." He was pointing to the contrast between the U.S., where the quarterly annualized growth rate stands in the 2% range from a year earlier, and Europe, including France, which is mired in crisis.
On concerns that AI is taking jobs, Manyika said: "We looked broadly at what is happening in the workplace, and in most cases we did not see a clear signal that employment is rising or falling, but we did see that pay is increasing in the fields most exposed to AI." He added: "That is very consistent with the idea that it can actually make those jobs more productive. Some people think that being exposed to AI means it's over for you, but in fact that's not the case."
Manyika said that from 2023 until now, as many as 200,000 jobs in the U.S. may have been eliminated by AI, but at least one million or more jobs have been spurred or newly created by AI, describing this as the essence of AI's impact on employment. "That said, we cannot be complacent," he said. "Even in a world where some jobs grow and some shrink, we have to address a range of skills training and transition issues right now."