For years, white-collar workers have been insulated from the perils of automation, sitting comfortably in their offices while reading news stories of blue-collar workers getting replaced by robots.
But emerging trends in artificial intelligence and automation are going to change that, according to entrepreneurs speaking at the think tank The New America Foundation Thursday.
One example of this new, super-capable form of artificial intelligence is StatSheet, a software program that takes sports statistics, integrates them with a human vocabulary and churns out news stories — all by itself — about baseball and football.
The application generates more than 15,000 articles a month and over the course of its nearly four-year lifespan, has created a million pages of news.
“It’s getting better every day,” said Robbie Allen, who invented StatSheet in 2007. “Within the next three to four years, it will be better than what a human can produce. And the reason for that is pretty much the foundation of computation: We can analyze and access significantly more data than one person can on their own.”
Complex computer programs aren’t just changing the business of writing — they’re also having profound effects on scientific discovery. Scientists in 2009 managed to design a program, called Eureka, which can figure out the laws of nature for itself, just by observing its surroundings.
The software works using what’s known as an “evolutionary” or “Darwinian” algorithm. Conventional approaches to AI design usually attempt to replicate human brain functions, said Michael Schmidt, one of the program’s architects. Problem is, that strategy usually produces a machine that’s subject to all the existing limitations of our own minds. Evolutionary algorithms are designed to overcome that hurdle by melding the acts of perception and analysis. Instead of matching data to pre-programmed mathematical formulas, Eureka teaches itself the formulas from scratch.
“You don’t have to design it from the top down,” said Schmidt. “It emerges from the bottom up. It’s not limited by our own understanding — that’s the whole idea.”
Schmidt believes scientists are on the verge of being left behind by their computers, which will seemingly be capable of leaping to conclusions while researchers spend most of their time puzzling over how their machines got there.
So what would an automated white-collar world look like?
Like most futuristic scenarios, this one has a dark side. Many of the thinking jobs that are expected to become automated in the near future are currently held by entry-level workers in office positions or college graduates. Replacing these individuals with machines could redefine middle-class incentive structures, said Michael Lind, policy director for the New America Foundation’s Economic Growth Program.
“Right now, most people going through university want jobs — that’s their goal. They’re not doing it out of pleasure of learning. If you don’t need that in terms of a job, why would people spend more money on education?” he said.
Extrapolating further, Lind explained that the shrinking incentive to get a college degree might discourage some people from investing in public education or neighborhood programs, which is why he believes government regulation will likely have a greater role to play in the future.
For his part, George Mason University economist Tyler Cowen thinks there will be an increasing demand for “feeling” over “thinking” jobs — service positions that reinforce human-to-human social interactions.
“We’ll basically end up hiring a lot of people to cheer us up,” he said. “A lot of jobs will be about motivation.”