Jan Hatzius predicted a soft landing in late 2022, back when many feared a recession was inevitable.

The Goldman Sachs chief economist made a name for himself by making the opposite call in 2008, warning that toxic mortgages would ignite a recession.

Now, Hatzius is offering a mostly optimistic forecast about another controversial theme in society today: artificial intelligence and what it means for the US economy.

Hatzius told CNN in an exclusive interview that he is very confident AI will significantly accelerate economic growth over time by making workers far more efficient.

“I see it as a productivity enhancer,” Hatzius said. “A large number of workers in the economy will become more productive. That is very, very likely.”

That productivity boost is expected to be so significant that it led Goldman Sachs last year to upgrade its long-term US gross domestic product (GDP) forecast.

AI chatbots can help workers brainstorm ideas, do research, write reports, build presentations, learn about new topics and identify patterns in vast troves of data. Even the Treasury Department and the IRS are turning to AI to fight financial crime and find tax cheats.

Of course, that’s not to say AI is perfect. It’s not. AI chatbots have been accused of bias and creating historically inaccurate images of people. AI tools are also known to sometimes “hallucinate” in a very believable way.

AI will ‘destroy’ some jobs, boost others

There is also the very real risk that AI will replace some workers. Generative AI is already able to write detailed emails, summarize dense books, craft witty advertisements and conjure up photorealistic images — all tasks that only humans were previously able to do.

“It will destroy employment in some areas,” Hatzius told CNN. “There will be parts of the labor market where tasks can be replaced. And to a degree, that is going to result in reduced employment there. But then you’ll also find other ways of innovating and creating more jobs somewhere else.”

White-collar workers are viewed as particularly exposed to this disruption.

Goldman Sachs previously estimated that as many as 300 million full-time jobs around the world could be automated in some way by generative AI.

Boosting GDP

Hatzius conceded it’s difficult to predict exactly which jobs will be destroyed and which ones will be saved.

“This is the story of economic growth and innovation for hundreds of years: You have an innovation that is basically labor saving and that reduces employment in some areas, but then boosts it in others,” he said. “How that balance is going to work out in the short term, it’s difficult to say. But where I’m much more confident is that it can significantly add to growth over time.”

Satyen Sangani, an economist and the CEO of data intelligence unicorn Alation, said the AI productivity boost will help offset stagnating labor forces in the United States and elsewhere.

“A lot of Boomers are retiring and labor is becoming more scarce. AI might be able to help slow the rate of decay in the labor force,” Sangani said.

For example, Sangani pointed to how AI chatbots can assist some customer support employees at understaffed hotels and medical professionals struggling to sift through complex medical records.

“These workers will be supplemented, not replaced, by AI,” Sangani said, though he added there are also places where AI will replace workers.

IMF warns AI could deepen inequality

Even if AI accelerates economic growth, there is no guarantee that everyone will benefit.

Earlier this year, the International Monetary Fund estimated that almost 40% of jobs around the world could be affected by AI and cautioned that this trend is likely to deepen inequality.

To fight the impact from AI, the IMF urged governments to build social safety nets and worker retraining programs.

In the meantime, investors continue to be captivated by the potential of AI. They are pouring billions of dollars into AI stocks, fueling a new gold rush on Wall Street.

But some are concerned the AI boom is overdone.

Jeremy Grantham, who predicted the dot-com crash in 2000 and the financial crisis in 2008, recently warned that AI is a bubble that could start to deflate.