How to know if a robot is about to steal your job
© John Lund/Blend Images/Corbis

Click here for free access to our latest coronavirus/COVID-19 research, commentary, and news.

Support nonprofit science journalism

Science’s extensive COVID-19 coverage is free to all readers. To support our nonprofit science journalism, please make a tax-deductible gift today.

How to know if a robot is about to steal your job

*For our full coverage of AAAS 2016, check out our meeting page.

WASHINGTON, D.C.—Drive a bus? You could be out of work in 25 years, thanks to robot drivers. Warehouse clerk? Your job could be gone, too. Supermarket cashier? Your gig could be lost to the machines—that is, if it hasn’t been already. Economists agree that technology is transforming the labor market, destroying entire categories of jobs while creating others. But they disagree heartily over the extent. Today, at at the annual meeting of AAAS (which publishes Science) here, a group of computer scientists and machine ethicists weighed in on that debate, saying that by the year 2045, automation could push down global employment—a measure known as the labor participation rate—to just 50%. Now, that number hovers just above 60%. Who are the losers? People in the middle of the skills curve, says Moshe Vardi, a computer scientist at Rice University in Houston, Texas. Those on the high end, including attorneys, doctors, and, yes, computer programmers, will still have skills (like creativity and the ability to contextualize) that even the most intelligent machines are unlikely to develop. Those on the low end, including food service workers, are paid such poor wages that the cost of automation wouldn’t be worth it. Meanwhile, folks like data entry clerks, hotel clerks, and almost anyone working in delivery or shipping are likely to suffer. How much? According to a recent report by the Oxford Martin School, automation will threaten 69% of the jobs in India and 77% of the jobs in China, compared with just 47% of jobs in the United States. Vardi is quick to point out that there is still no good way of comprehensively measuring that kind of risk. But a transformation is still taking place. “We need to start thinking very seriously: What will humans do when machines can do almost everything?”