This is something you read a lot about these days: Robots are coming for your job; AI is driving human employment into extinction; the X jobs AI will take first. The articles are often accompanied with feature images of penniless humans and rich robots, robots enslaving humans, and worse.
To be fair, the warnings are founded in solid facts. It’s undeniable that artificial intelligence is transforming the employment landscape, just like every other technological revolution. However, the picture that is being portrayed is often too somber and overblown, heralding a near-future where human labor has become non-existent.
Part of the hype is due to the huge progress that tech companies have made in robotics and artificial intelligence, enabling computers to accomplish feats that previously seemed impossible. But the truth is that we still have a ways to go before we bring technological employment down on ourselves. And there’s a lot we can do to make sure we don’t create a robot apocalypse, Skynet, Genysis, or whatever dark name a dystopian AI future may be called.
So while speculating on the long-term disruption of employment, we should also try to better understand and prepare for the short-term displacement of jobs by artificial intelligence.
The main driving factor behind the recent AI revolution is machine learning. Machine learning algorithms function by ingesting huge sets of data and inferring patterns, which they use to make classifications, perform tasks and predict outcomes. This makes them especially useful for tasks that can’t be programmed in the traditional rule-based manner. ML depends on huge amounts of data and computation power, both of which are becoming available in increasing abundance and at low costs.
Machine learning and deep learning, its more advanced subset, are behind the automation of many routine jobs, especially cognitive tasks such as sales, accounting and customer service.
However, this does not mean that jobs are going extinct.
In a feature published in Wired, James Surowiecki explains that if automation was transforming the economy, two things should be apparent: an aggregate rise in productivity and a dwindling number of jobs. However, the figures tell a different story. For instance, in the U.S., where technological unemployment is hotly debated, employment churn is at an historical low and median job tenure has lengthened.
“It’s logical to conclude that if a job can be…