AI optimists are convinced that the new technology is going to propel the global economy into extreme levels of economic growth, believing that GDP growth could explode to an unprecedented 20-30 percent annually in the near future. Researchers at Epoch AI estimate that if AI is able to automate 30 percent of tasks, GDP growth would be 20 percent.
Such growth would be enabled by self-improving AI. A widely circulated paper called “AI 2027” claims that AI could become super-intelligent in the next few years, and when increasingly intelligent models develop and train future models instead of human developers, AI models will begin to improve at an exponentially increasing pace. According to the paper, this could lead to unimaginable technological innovation, world peace, and one world government by 2030—or a Terminator-like human extinction event that could be the end result of moving too fast with AI models that don’t have humanity’s best interests in mind.
“My experience, talking with a lot of those folks, is that the real difference [in Silicon Valley] is between people who think this is all going to happen by 2027 and those who think it’s going to happen by 2035,” James Pethokoukis, a senior fellow at the American Enterprise Institute, told TMD. “The cautious people are like, ‘It’ll take a decade.’ That’s not like the Fed or [Congressional Budget Office] forecasts.” The Fed, for example, predicts annual GDP growth of roughly 2 percent over the longer run.
Economists, on the other hand—while acknowledging AI’s potential to increase productivity—are less convinced that AI will drastically shift the economic landscape. At least for now.
JP Morgan anticipates an 8-9 percent boost to GDP over the course of the next decade—a very different prediction than 30 percent annual growth some in the tech world are promising. Goldman Sachs is more bullish on AI, estimating a 15 percent GDP increase over the next decade—strong growth, but not “world peace” levels. Other economists are less optimistic. MIT professor Daron Acemoglu predicts that, in the next decade, AI will only raise GDP by roughly 1 percent, assuming that the technology is only able to automate 5 percent of tasks.
While AI is already drawing billions of dollars of investment and is driving much of the stock market’s growth, economists are looking at different indicators to determine how much the technology will boost the overall economy. Whether businesses are adopting AI, working it into their business operations, and observing improved worker productivity matters much more to overall GDP growth. These are the metrics economists have their eyes on, and so far, they aren’t matching up with Silicon Valley’s lofty predictions.
“The adoption of this technology has been fast compared to some technologies of the past, but the general economist answer is that something can be really amazing in the lab, but it takes businesses a long time to figure out how to effectively use these technologies,” Pethokoukis said.
Right now, businesses are keen to adopt AI, but they’ve yet to see significant company-wide improvement. McKinsey & Company analysts assessed that 78 percent of companies are using AI in at least one business function, but only 19 percent of C-Suite executives reported a more than 5 percent increase in revenue, and 74 percent said AI adoption had either increased or not changed their costs. Many firms may be looking longer-term. According to McKinsey, 51 percent of executives say they expect AI to increase revenue by more than 5 percent in the next three years.
“I think there is every indication this is a very important technology and what they call a ‘general-purpose technology,’ meaning it can be used in lots of different industries. … It’s more like electricity or the internet. If you assume that it’s kind of like the internet, then it will eventually make people more productive and that will show up in the data. That’s sort of the cautious, non-Silicon Valley expectation,” Pethokoukis said. “The Silicon Valley expectation can be science-fictional sometimes.”
But while the technology could lead to growth in economic output, many are concerned that AI will soon shut humans out of the job market. And there’s some anecdotal evidence that it may already be happening. The Wall Street Journal reported this week that AI is taking over entry-level jobs that would otherwise go to recent college graduates. The New York Federal Reserve reported that the unemployment rate for recent college graduates currently sits at 5.8 percent, its highest rate since 2021.
But there are likely other factors at play. “We’re still a little ways away from seeing AI showing up truly in labor market data. There’s a lot of anecdotes about this happening, about entry-level jobs disappearing and it being a really hard time for college graduates, but there are a lot of additional factors that are going on in the economy that make it really difficult to say whether this is truly AI or not,” Tobias Sytsma, an economist at RAND, told TMD. “Even just business uncertainty can reduce hiring in different industries.”
Looking at whether entry-level job shortages persist over the next few years will paint a clearer picture as to whether AI is currently disrupting the workforce at a large scale. “That could be a sign of what’s to come, or it could just be we’re seeing a bunch of things in the economy and we’re trying to assign causality to something that maybe is just correlational,” Sytsma said.
Most industries are also not using AI at a large enough scale to cause significant labor market disruption yet, either. “Business adoption rates are still very, very low,” Pethokoukis said. “It’d be surprising if we had this huge impact.” According to Goldman Sachs research, sectors more exposed to AI have shown few statistically significant differences in key labor market indicators—such as job growth, wage gains, unemployment rates, and layoff rates—from sectors that are less exposed.
Silicon Valley, however, is prepared for a much faster change. The CEO of the AI company Anthropic said he anticipates that unemployment could jump to 20 percent in the next one to five years. Such levels, coupled with extremely high GDP, could reshape the entire economy, pushing people to own capital rather than hold a job. Investments would pay off in a booming, but workerless, economy—holding a job would no longer be the most reliable way to make money.
But key bottlenecks will likely slow AI’s effect on the economy. Companies building data centers are struggling to meet the energy demand, and there are already arguments over who is going to pay for the inevitable upgrades to aging U.S. power infrastructure. “We’re having trouble building in a quick, budgetary way, all these new power sources,” Pethokoukis said. “All those kinds of bottlenecks don’t go away just because you have an AI model that can pass every benchmark or Math Olympics.”
Businesses also need to trust AI for it to be adopted widely. The technology is far from perfect, as demonstrated by Grok of xAI’s recent dalliance with hate speech. If business leaders believe AI isn’t ready for use in higher-risk decisions, it could limit adoption. Companies still have not seen significant revenue boosts or cost reductions from AI, either. If the technology continues to be financially unviable, AI’s impact on the economy will be muted. It doesn’t mean Silicon Valley’s hopes are necessarily wrong, but radical economic changes might be farther off than tech CEOs might envision.
“There are a lot of vibes-based forecasts. It doesn’t come from history,” Pethokoukis said. “For now, I’ll go with economic history, but I’m open to that the vibes might be right.”
