Reports of the death of the job as a result of AI are greatly exaggerated. There are just too few things that robots and artificial intelligences can do better than humans at this point.
That’s according to a recent interview in Wired with Erik Brynjolfsson, PhD, director of the MIT Initiative on the Digital Economy, who argues that as long as companies are prepared to restructure and retrain as necessary, AI will be a boon, rather than a menace, at work.
In the interview Dr. Brynjolfsson explains how, particularly in healthcare, AI and robotics should be seen as a tool not only to improve accuracy, but also to free up humans to complete other, more in-depth tasks.
“I subscribe to the narrative that mass job replacement isn't here,” Brynjolfsson claimed. “What is imminent is the replacement of parts of jobs through AI but also through robotics.”
He cites the example of a machine learning algorithm analyzing medical images in radiology that “might be 97 percent accurate and a human might be 95 percent accurate, and you might think, OK, have the machine do it. You’re better off having the machine do it and then a human check it afterwards. Then you go from 97 percent to 99 percent accuracy because humans and machines make different kinds of mistakes.”
Beyond that, the human radiologist will do a much more efficient job on their own when it comes to comforting patients and coordinating with other physicians — essentially, any task requiring human connection.
“Our brains are wired to react emotionally to other humans,” he points out. “Humans just have a comparative advantage at connecting with each other. We're very far from Westworld, and even there the robots weren't always that convincing. That's not where we are or will be anytime soon. I think this is great news, because for most of us, the parts of our job that involve creativity and connecting with others are the parts we like best. The part we don't like is repetitively lifting heavy boxes, and that's exactly what machines are really good at.”
Successfully incorporating AI into the workplace, then, will require some restructuring and, if necessary, retraining to ensure humans are in positions where they are most needed.
"I think it's a little bit of a lazy mindset to look at a business process or a job and just sort of say, OK, how can a machine do that whole thing? That's rarely the right answer," Dr. Brynjolfsson said. "Usually the right answer requires a little more creativity, which is how can we redesign the process so parts of it can be done by a machine really effectively and other parts are done by a human really effectively, and they fit together in a new way."