“The current systems are actually not very good at all at doing whole jobs,” he said on Wednesday at an Economic Times event in India. “They’re very good at doing tasks.” 

Using the example of a computer programmer, he said, “the nature of the job shifts to, you kind of manage a team of extremely, extremely junior developers that can only do one one-minute task at a time.”

Last week, HP CEO Enrique Lores said A.I. will significantly change how employees use their time. “Many activities that today take one day to get done are going to take seconds,” he told Yahoo Finance. “And it’s going to be really shifting the world from doing things to interpreting things, to really working with the output.” 

Of course, the A.I. tools will get better.

“Someday they’ll do 10-minute tasks, and then they’ll do an hour task,” Altman continued. “But you’ll still have to think about, ‘How is this all going to fit together? What do I want to build?’ And, you know, maybe it eventually learns that, too.” 

But for now, Altman said, A.I. is more likely making workers “dramatically more efficient” than replacing them.

He also veered toward an argument recently made by IBM CEO Arvind Krishna and former Google CEO Eric Schmidt—essentially that with declining populations in much of the world, there won’t be enough people to do all the jobs that need doing, and A.I. can help by taking on some of the work.

“In aggregate, all the demographics say there’s going to be a shortage of humans for jobs—literally too many jobs and not enough people for at least the next 30 years,” Schmidt said at the Wall Street Journal’s CEO Council Summit last month.

Given that, having people perform tasks that A.I. could do is “not an option,” Krishna said in a recent Nikkei interview. “We are going to need technology to do some of the mundane work so that people can do higher-value work.”

Altman added his own thoughts on this, saying: “Maybe the problem is we don’t have nearly enough people to do all the jobs we want to.”

He continued, “There is such a demand overhang in most places. If we can overnight make the world create 3X more software because we make every software developer three times more efficient, that is not nearly enough—that does not nearly fulfill the demand the world has for software. And I think we’ll see that in many other places.”

But the fear of A.I. stealing jobs is widespread. Altman’s onstage interviewer in India—Satyan Gajwani, vice chairman of media company Times Internet—noted: “The natural fear is that A.I. is going to make us redundant, particularly in markets like India, where we have so much of a workforce, and a lot of it oftentimes is doing somewhat rote work. Should we be worried about this?”

Altman replied that “every technological revolution leads to job change” and acknowledged that “some jobs will go away” owing to the ability of A.I. to do certain tasks. But he insisted that “there will be new, better jobs,” even if “they’re difficult to imagine as we sit here and dream about what the future is going to look like.”

Joseph Fuller, a Harvard Business School management professor, recently told Fortune that jobs involving rote duties will likely disappear. He gave the example of a routine contract lawyer who writes standard submissions, or “someone who does the reading or summarization of business books to send out 20-page summaries, because A.I. is really good at summarization already.” 

Altman remains upbeat about our future with A.I. “I think the world will get way wealthier, you’ll have a productivity boom, and we will find a lot of new things to do,” he said.

But he also warned that the speed of change with A.I. could cause problems. “In two generations, we can kind of adapt to any amount of labor market change…The thing that might be different about this is the speed with which it could happen. And I think it will require a change to the socioeconomic contract and the way governments think about this if it happens at a very fast pace.”

Goldman Sachs estimated in March that while A.I. might lead to new jobs and a productivity boom, it could also expose the equivalent of 300 million full-time jobs to automation.

In March on the Lex Fridman podcast, Altman spoke of efforts by himself and OpenAI to support universal basic income (UBI) schemes. The idea behind UBI is that governments would give every person a tax-free, flat amount of income, regardless of their wealth or employment status. Altman told Fridman that UBI could serve as a “cushion throughout a dramatic transition” brought on by artificial intelligence.

In 2019, Altman cofounded Worldcoin, a crypto project that could help distribute UBI payments after verifying recipients’ identities via an eyeball-scanning orb-shaped device.

Altman believes we need to start thinking about UBI now—even though today’s A.I. tools are better at replacing tasks than jobs—because of how quickly the technology is improving.

“Systems in the near future will be dramatically different,” he said yesterday. “None of the current systems matter. We are on an exponential curve, really. ChatGPT 3.5 is like the old grayscale Nokia phone, and the ‘iPhone 14 of A.I.’ will be coming soon.”