Artificial intelligence probably won't kill you, but it could take your job
Computer scientist Zachary Chase Lipton hates the term artificial intelligence, which he says gives people the wrong idea. He prefers to call it machine learning.
Source: Don Pittis
Lipton, who used to earn his living as a musician, would be a dream example for Prime Minister Justin Trudeau and Finance Minister Bill Morneau, who are hoping to retrain Canada's underemployed to serve a resurgent artificial intelligence industry.
But the AI whiz says retraining may not be enough to prevent a wave of social disruption when the tech economy leaves the poor and middle class behind.
Lipton is not one to scoff at the opportunities offered by retraining.
"My employment prospects are considerably greater owing to machine learning than they were five years ago when I was playing saxophone in New York for 50 bucks and drinks tickets," says Lipton, on the phone from Palo Alto, Calif.
After deciding in his late 20s to switch from playing the sax to playing with computers, Lipton is now a paid adviser to the giant online retailer Amazon and has just been head-hunted as a professor by Carnegie Mellon University. He hasn't even completed his PhD.
Zachary Chase Lipton, Computer Scientist
Only five years ago Zachary Chase Lipton was playing jazz, but even though he now works in artificial intelligence for Amazon and has accepted a professorship, 'technological unemployment' is his biggest concern. (Point14 Records)
Also last week, Lipton made a splash in tech circles with his blog post, The AI Misinformation Epidemic, one of a series of articles he is planning under the rubric Approximately Correct.
Future filled with possibility
In Canada there has been an outbreak of coverage on artificial intelligence as Canadian governments and businesses invest in AI research, including the new non-profit Vector Institute.
"By encouraging cutting-edge technology like AI while at the same time creating a culture of lifelong learning, we will be with Canadians every step of the way as they lead us into a future filled with possibility," said Morneau at last week's institute launch.
Lipton agrees that the technology is valuable in many areas, especially in such things as improving the efficiency of medical systems. But he worries that so much attention by the popular media on what he calls "artificial general intelligence" and intelligent killer robots means critics mistakenly suggest that AI is dangerously powerful.
"Another worry is that AI is not that strong, but it's good enough to be convenient," says Lipton. The terms AI and "neural networks" lead the misinformed to imagine these systems replicate human intelligence, something he says won't happen for a long time, if ever.
He offers the example of a Facebook algorithm that throws up fake news over important news. He has seen ostensibly impartial AI systems, intended to predict prison recidivism rates, which embed racial discrimination.
Taking humans out of the equation
But like many other Silicon Valley computer prodigies, his biggest worry has little to do with mathematics or science. Lipton, who studied economics as an undergraduate, fears that AI is fundamentally transforming the economy to take humans out of the equation.
Traditional economics tells us that capital and labour are complementary. More ovens (capital) are no good without more bakers (labour). But we are at a watershed where that may have changed.
"I think now with automation we'll be investing in capital for the purpose of replacing labour," says Lipton.
A survey last year of CEOs by the executive search firm Korn Ferry showed 44 per cent said "robotics, automation and artificial intelligence (AI) will make people 'largely irrelevant' in the future of work."
That kind of thinking worries Richard Mueller, head of economics at Alberta's University of Lethbridge. He observes that everyone imagines driverless cars and trucks are coming in the distant future, including his brother-in-law, who drives a giant truck at the coal mine in Elkford, B.C. But within the controlled conditions of private mine sites, driverless trucks are already being used around the world.
"Those guys are ripe for replacement almost immediately," says Mueller, who has given a series of lectures on the future of work in the automation age.
No one can predict the future, and so far the unemployment rates in Canada and the U.S. have been falling. But a new study from the U.S. National Bureau of Economic Research has observed that automation is winning the race for U.S. jobs.
While Both Mueller and Lipton support the idea of retraining, they fear that many of those disinclined to technical skills may never be able to earn a good living in a market economy.
Mueller worries that as AI improves and gets cheaper, many of the jobs left for humans will be those so badly paid they are not worth replacing with a machine.
Lipton, who thinks the solution may ultimately require a redistribution of wealth from capital to labour, thinks that the biggest problem could be helping those whose jobs have been stolen by smart machines to find meaningful lives.
"The one thing we shouldn't automate is music," quips the sax man. "We'll teach everyone to play jazz. Your purpose in life is to swing."
Follow Don on Twitter @don_pittis