Kevin Carmichael: Don’t fear the robots, people

Article content

History is imperfect. Most of the original stories were written by conquerers, so they’re one-sided. Modern historians have done hero’s work trying to correct the record by digging into primary sources, but myths persist: There is only so much data a professor and a team of research assistants can process, even when ancient ledgers and old diaries have been digitized.

Revisionist history is about to be turbo-charged thanks to companies such as Amazon.com Inc. Some researchers are shivering with anticipation. Others probably are trembling with fear for their jobs. The mention of Amazon will have given away the cause of such strong feelings: The robots have escaped the factory floor and are looking for new things to do.

The dichotomy captures the uneasiness with which humanity confronts big technological change. Forget the Luddites, the most famous technophobes. In 1989, Statistics Canada asked 9,338 people, “Do you agree or disagree that, on balance, computers and automation will create more jobs than they will eliminate?” Twenty-four per cent of respondents disagreed strongly, 28 per cent said they somewhat disagreed, and 15 per cent had no opinion.

Article content

The way we choose to confront the tech revolution will set the parameters for the kind of economy we have for decades to come

So, more than half the population was less than enthusiastic about what computers would mean for the labour market. Their pessimism was misplaced. The services side of the economy boomed and employment grew by about six million positions between 1989 and 2019. The percentage of the working age population with a job held constant at about 62 per cent.

The way we choose to confront the current technological revolution likely will set the parameters for the kind of economy we have for the next many decades. Some of today’s pessimists favour taxing algorithms and robots to keep employers from abandoning most of their humans. Others reckon that technology will make most of us better and richer, like it always has. One of Stephen Poloz’s final messages as Bank of Canada governor was that he thought the shock of the COVID-19 crisis and the acceleration of digital work and commerce would spark a renaissance in entrepreneurship. “It will happen in spades,” he said in an interview in May 2020.

So the stakes are high. Those who get the shift to a digital economy right will emerge as winners, and those who choose poorly will become the takers. The latter group could turn out to be perfectly happy, but it will be relatively poorer. That probably matters, if only because the winners will have an easier time paying the debts they accumulated fighting the COVID-19 crisis.

I recently spoke with Nicholas Zammit, an economic historian at the University of Toronto’s Mississauga campus, his research assistant, Dev’Roux Maharaj; and Carol Kennet, an executive at Amazon Web Services Inc., the biggest provider of cloud-based computer services.

Article content

Zammit and Maharaj are exactly the sort of people you want sifting through old records: nerds in the most positive sense. Kennet is more complicated. She’s either the holder of the keys to the kingdom or a stalking horse of the robot apocalypse, depending on your point of view. Her title is education and digital innovation lead at the University of British Columbia Cloud Innovation Centre, a two-year-old partnership between AWS and UBC that pairs the school’s brainpower with Amazon’s computing power.

The gap between the humans and the machines is shrinking

The gap between the humans and the machines is shrinking. Perhaps the most powerful tool that Amazon brings to the project is Textract, software that uses artificial intelligence (AI) to “read” raw data, extracting the relevant bits for analysis. The partnership with UBC is characterized as benevolent. The Cloud Innovation Centre accepts applications from anyone with an idea, although the emphasis is on solving puzzles related to human health. The research is open source, so Amazon isn’t out to steal the ideas of bright academics. Still, it probably wouldn’t mind if its efforts led to some business for AWS.

“We want to show ways these technologies could be used,” Kennet said.

Amazon and some of the other internet behemoths will own the future because they own most of the cloud. Businesses such as AWS, Microsoft Corp.’s Azure, and Google’s cloud offerings have become much more than data warehouses. They have built the foundation on which the digital economy will sit and, so far, there’s just enough competition to spur those companies to make tools like AWS’s Textrack available at affordable prices. That’s an opportunity for entrepreneurs, creatives and academics that didn’t exist until very recently.

Article content

Consider Zammit. His current project is an attempt to put an economic price on the cost of war by measuring what happened to international trade when Canada and other Dominions joined the First World War. The research was gruelling until he and Maharaj convinced the Cloud Innovation Centre to back their work. Together, they developed a tool that scans digital versions of primary sources and transfers the relevant data to a “very, very big” spreadsheet.

Before, it would take one person working full time for four months to analyze one year’s worth of data. “We have about 30 to 40 years of data to analyze,” said Maharaj. “Do the math. That’s a lot of time. And we have to pay those people. You need to pay those people and that’s a lot of money.”

Now, neither time nor money are important barriers. One person working about 10 hours per week can collect and organize a year’s worth of data in half the time (two months). What’s more, that person is collecting and organizing a richer set of data because AWS’s AI is mining information that its human masters had abandoned.

“We’re going to have so much more depth of what we can see going on in, say, the 19th century,” said Zammit . “Before, we were ignoring a lot of cells on the spreadsheet, because we assumed we’d never have time to enter all of the data. Now we’re going to have everything.”

Should we fear such a future? Zammit’s research will lead to a richer understanding of what drove trade patterns before and during the War, which could in turn either support our current notions about the merits of freer trade or debunk widely held biases about the harm caused by protectionism. He’ll be able to do more, faster.

What of those unemployed research assistants? They’ll be fine. Research suggests that fancy degrees offer decent protection from robots. They’ll just need to think for themselves rather than wait for professor to tell them what to do.

• Email: | Twitter: carmichaelkevin