A.I. could automate 30% of STEM workers’ hours by 2030

The folks at McKinsey Global Institute, whose study prompted my post in 2017, have attempted to attach some new numbers to the change. Generative A.I., they write in a new report out this morning, can be used “to write code, design products, create marketing content and strategies, streamline operations, analyze documents, provide customer service, and even accelerate scientific discovery.” Before generative A.I., their research estimated automation could take over tasks accounting for 21.5% of the hours worked in the U.S. economy by 2030. With the new technology, that number jumps to 29.5%.

The biggest change is for STEM professionals, where automation potential by 2030 jumps from 14% of work hours to 30% of work hours. Similar big jumps occur for education and training work, creative and arts management work, and business and legal automation. The study does not conclude that generative A.I. will lead to a drop in jobs, but rather that those jobs, and the mix of activities they involve, will change dramatically. “Just under 12 million workers may need to find new occupations by 2030,” the report says. The idea is to start preparing now, beefing up training programs both inside and across companies.

Separately, Fortune yesterday released the inaugural Fortune China 500 list. Some companies on it—like No. 1, State Grid, the government-owned electric utility that had $530 billion in revenues last year—already rank high on the Fortune Global 500 list (out in its latest iteration next week). Others, like No. 500—the Inner Mongolia Mengdian Huaneng Thermal Power company, with revenues of $3.4 billion—may be less familiar. But overall, the list provides an interesting snapshot of the world’s second-largest economy.

Alan Murray, Nicholas Gordon