“People mistake productivity with job displacement,” Krishna said onstage. 

He noted that, as IBM phased out a few hundred back-office HR roles over three to four years, as it added headcount in software engineering and sales roles. “The increase was like 8,000,” Krishna said. “The decrease was like 800.” 

Krishna specified that employees weren’t let go as a result of this transition—rather, certain roles were not backfilled when they opened up. 

But there’s no question that “repetitive white-collar jobs” will be affected by the technology, Krishna said, echoing a point he has raised in the past.

“The first thing you can automate is a repetitive, white-collar job,” he said on Tuesday. But while AI could take over 10% to 20% of “lower level tasks,” he predicted it wouldn’t take a person’s job altogether, because no one’s job is composed entirely of these sorts of tasks, he says. He expects his programmers to get 30% more productive thanks to the technology. “I don’t intend to get rid of a single one,” he said. “I’ll get more.” 

All of which is a boon for developed countries, where Krishna sees an ongoing labor shortage in the coming years. During his onstage interview at the CEO Initiative, Krishna drew a contrast between the scarce labor market in the developed world, which will be in dire need of supercharging productivity, and the developing world, which will have lots of ready and willing workers to fill open jobs. That’s similar to a point he made in a May interview with CNBC, saying successful implementation of AI was critical to maintaining the current quality of life the developing world enjoys. 

“Population is flat or, in the worst case, declining,” Krishna said at the time. “So you need to get productivity, otherwise, quality of life is going to fall. And AI is the only answer we got.” 

The increase in productivity will allow companies like IBM to hire more people, not fewer, because it will allow them to produce more goods and services, which will then need to be taken to market. He gave the example of the rise of the smartphone, which unexpectedly created a booming market for software designers. “In 1995 no one thought there would be five million web designers—there are,” Krishna said. His fellow panelist, Qualcomm CEO Cristiano Amon, gave the example of Uber, which created an entirely new use for the smartphone that hadn’t been previously considered but became rather ubiquitous in consumers’ lives. 

At IBM, AI has already been integrated into some of the company’s most critical work. About three-quarters of chip design at the company is done by AI, according to Krishna. In July, Reuters reported that IBM was considering using its own chips to power its cloud computing service to lower costs. 

‘China is advanced in AI, let’s not have any illusions about that’

The conversation then veered toward another hot-button issue in AI: the ongoing arms race with China. 

“China is advanced in AI,” Krishna said. “Let’s not have any illusions about that.” 

Although Krishna’s view, which he called a “little bit contrarian,” is that the U.S. is leading in AI, the size of that lead is debatable, he said: “Is the U.S. six months ahead or two years ahead?”

China does have some advantages compared to the U.S.—most notably, a government comfortable issuing dictats on AI development in order to support its own aims. China, Krishna says, has the “ability to use large private datasets,” something U.S. firms can’t do because of privacy and other laws. The government in China is also more open to exploring uses of A.I. the U.S. might consider undemocratic. All of which will result in a fracturing for the application of AI. “As there’s more and more of a tech war, there’s going to be more and more of a development that’s splintered across the two nations,” Krishna said. 

Amon, who noted that the U.S. remains the global leader in chip design, predicted that the U.S. could become the dominant force in AI and all the tangential industries needed to support its development. 

“The most important thing we can do is make sure the US remains the number one place to attract talent because there’s a finite number of individuals that can help create AI and those technologies,” he said. 

Krishna then reminded the audience that the U.S. has been the cradle of AI technology. “[That] generative AI came out of places like Stanford, MIT, Google, and Facebook is a fact that I think we should all take pride in.” 

IBM did not immediately respond to a request for comment.