Machine Learning Could Be As Important To Future Of Work As Electricity Once Was, Expert Says

Jan 8, 2018

The future of work will hinge on machine learning technology, a type of artificial intelligence that improves performance with experience, according to Carnegie Mellon University's Tom Mitchell. 

In a report published recently in the journal Science, Mitchell, a machine learning professor, said the technology will transform the economy as much as the steam engine and electricity did in the past. This will result in "good and bad impacts" on the future of work.

"Jobs won't so much disappear, they'll just take on a new shape and form that pushes more toward the human doing non-routine parts of a job," Mitchell said. "They'll do more of the interaction with people part."

Computers and doctors are already working together in diagnosing diseases in hospitals, Mitchell said. Late last year, UPMC reported a computer program that uses machine learning helped the hospital researchers detect more cases of acute kidney failure, saving lives. 

But Mitchell said some jobs that involve one repetitive task, such as toll booth operators, are more likely to be automated.

The future of work will likely affect the future of education, according to Mitchell. He said education will focus on the skills machines miss, such as emotional intelligence.

"The key insight here is to realize that most jobs actually involve multiple tasks," he said. "And when you think of automation, it's usually about a single task."

Machine learning is already widely used for credit card fraud detection and financial market analysis.