“They will learn from the workflow they are dealing with,” Autor says. “Often people are busy working with a tool, and the tool will learn from that interaction.”
Whether you train an AI tool directly by interacting with it all day, or whether the data you produce as you work is simply used to create an AI program that can do the work you do: there are several ways in which an employee could accidentally train an AI program to replace him. Even if the program doesn’t ultimately prove to be incredibly effective, many companies might be happy with an AI program that’s good enough because it doesn’t require a salary and benefits.
“I think there are a lot of discretionary white-collar jobs where you use a combination of hard and soft information and try to make sophisticated decisions,” Autor says. “People aren’t very good at that, machines aren’t very good at that, but machines can probably be just as good as humans.”
Autor says he doesn’t see a “labor market apocalypse” coming. Many workers will not be completely replaced, but rather their jobs will simply be changed by AI, Autor says, while some workers will certainly be laid off by advances in AI. The problem there, he says, is what happens to those workers after they can no longer find a good-paying job with the education and skills they have.
“It’s not like we’re going to be out of work. It’s much more the case that people do something they’re good at, and that thing disappears. And then they end up doing some kind of generic activity that everyone is good at, which means it yields very little: serving food, cleaning, security, driving,” says Autor. “These are low-paying activities.”
Once someone is automated out of a high-paying job, they can slip through the cracks. Autor says we’ve seen this happen in the past.
“The erosion of manufacturing and office work over the last forty years has certainly put downward pressure on the wages of people who would do that kind of work, and it’s not because they’re now doing it at lower wages. It’s because they don’t do it,” says Autor.
Frey says politicians will have to provide solutions to those who fall through the cracks to avoid the destabilization of the economy and society. That would likely include offering social safety net programs to those affected. Frey has written extensively about the consequences of the first industrial revolution, and he says there are lessons to be learned. In Great Britain, for example, there was a program called the Poor Laws, which provided financial assistance to people harmed by automation.
“What you then see is a lot of social unrest. For a large part of the population, wages are stagnating or falling. You have riots,” says Frey. “If you look at the places where the Poor Laws were more generous, there was less social unrest and less unrest. Using social security systems to compensate those who have suffered is something we have been doing for a long time and we must continue to do so.”