|
Written by Kieran Delamont, Associate Editor, London Inc. | |
AI
Is Claude ... Canadian?
AI models are built by U.S. tech giants trained on American-dominated data. So why are they biased toward Canadian-like answers?
| | |
SPIRITUALLY, YOU MIGHT say that the big AI companies are quintessentially American: moving fast, breaking things, pursuing growth at a rapid clip and deciding it’s better to beg for forgiveness than to ask for permission (the job market thanks them for this).
The flagship chatbots, however, may not be so American, at least not in terms of their values and philosophical tendencies. When it comes to those, Claude is, in a lovely display of nominative determinism, rather Canadian.
At least, that’s the argument from Transformer Lab, an AI research shop that set out to test the assumption that U.S.-built AI models would have U.S.-centric worldviews. By testing five major LLM models against survey responses from around 6,600 North Americans, it found that few of the models were all that American in their responses.
“In 69 per cent of statistically significant comparisons, a model’s default answer distribution landed closer to Canadian public opinion than American,” Transformer Lab’s report reads. “Every single model leaned Canadian.” The effect was strongest when the models were prompted on topics related to government, institutions and national pride, were most even on questions of religion and were most American when asked questions about interpersonal trust.
“When a Silicon Valley company tries to make a model that’s perfectly polite, universally tolerant, highly trusting of institutions, really afraid to encourage any extremism, you end up accidentally building a Canadian,” said Transformer Lab co-founder Ali Asaria, speaking to The Logic. “Maybe what an American most wants out of its best employee is a Canadian.”
The researchers note that the data doesn’t say exactly why this is the case, but their working theory is that the cultural biases of the people who train the language models is what matters.
“The explanation we find most compelling is that we are looking at the values of the people who aligned these models, not the raw internet text underneath them,” they wrote. “The annotators and developers steering that process tend to be highly educated tech workers whose collective preferences — tolerance, diplomacy, trust in institutions — may map closer to average Canadian opinion than to the polarized American public.”
| |
WORKFORCE
The job-hugging trend is cracking. Why a quarter of Canadian employees are now looking to leave
Despite a seemingly frozen labour market, subtle shifts are tipping the balance of power back to workers
| | |
AFTER A TURBULENT couple years in the immediate post-Covid environment, the last 18 to 24 months in the job market has been dominated by the ‘great stay/job hugging’ dynamic, with workers clinging to their existing jobs rather than setting out to find new ones. It’s been ‘low hire, low fire’ since mid-2024, and low turnover and churn in the job market has reflected that.
Well, here comes the pendulum again. It seems Canadians are getting itchy to change jobs again, with nearly one in four planning to leave, a number that rises to four in ten when looking at Gen Z, according to a new survey from Randstad Canada.
But what is slightly different compared to the post-pandemic era of job switching (where hopping from spot to spot was largely about improving your salary), is that this time around a lot of the restlessness seems to be related to working conditions and flexibility.
“Roughly eight in 10 workers say flexibility, time off and health and wellness benefits are important,” wrote Jim Wilson at HR Reporter, noting that younger workers are placing a higher premium on work-life balance than older workers right now. Against the backdrop of RTO and a culture of productivity-maxxing, people are ready to look elsewhere.
“This year’s data reminds us that the foundation of the employer-employee relationship hasn’t fundamentally changed: workers want to be fairly compensated and able to live their lives outside of work,” said Marie-Eve Robitaille, division president of professional talent solutions at Randstad Canada. “What has changed is how quickly they act when they feel those needs aren’t being met.”
Last time around, a lot of companies struggled to navigate an employment market with more restlessness; complaints about flighty employees jumping ship were common, circa 2022 and 2023. But Marie-Pier Bedard, also of Randstad Canada, told the Financial Post that companies don’t have to struggle with a greater desire to leave.
“Employers should expect a highly competitive hiring season where candidates aren’t just daydreaming about finding something better — they are actively ready to jump ship,” she said. “Talent is going to gravitate toward companies that clearly deliver on what they value most. Navigating this doesn’t require employers to completely reinvent the wheel.”
| |
| |
Terry Talk: Spain’s World Cup secret? Succession planning
| When Spain won yesterday’s World Cup, the real story wasn’t about a superstar carrying the team. It was about years of intentional talent development. The coach spent nearly a decade building players, creating depth, trust and a system that could withstand pressure. Too many organizations search for the next expert, rainmaker or gamechanger. But sustainable success rarely comes from one person. It comes from developing talent, building strong systems and creating a team that works together when the stakes are highest. In this Terry Talk, Ahria Consulting president & CEO Terry Gillis explores what Spain’s victory can teach leaders about succession planning and talent development, and how ordinary people can accomplish extraordinary things when they’re part of something bigger than themselves. | | | |
PRODUCTIVITY
Giving credit where it’s due
Bosses want you to use AI. Then they credit the AI
| | |
THE RISE OF AI agents in the workplace has started posing a new question: when work goes well, who — or what — should get the credit?
A story in Business Insider plumbed this quandary, and found that in many corners of the professional world this is a growing problem. “White-collar workers are caught in a nasty pickle,” wrote Shubham Agarwal. “Many employees have begun hiding their AI usage and wonder how much credit, if any, they should give it for their efforts. To credit or not to credit AI can feel like as vital a question as to be or not to be.”
There’s hardly one way to use AI, and work styles using AI tools differ from person to person, yet research from Christoph Riedl cited by Business Insider points out there are trends in how work is evaluated when humans work alongside AI.
“Managers consistently devalued workers’ contributions to projects when workers revealed AI had assisted them,” they said. It seems many managers assumed, often incorrectly, that human input was minimal.
Some workers view the question differently — you don’t credit Excel for good accounting, for instance, or credit Microsoft Word for writing a good paragraph. Others still wonder if that’s simplifying it a bit too much.
“The issue becomes whether the apparent intellectual contribution can still be attributed to the human author or agency was largely outsourced to the machine,” said management professor Oliver Schilke.
Others still think it’s the wrong question entirely, and that what matters isn’t how the work was produced but “whether the person responsible for it can defend it, improve it and be held accountable when it fails.”
Some experts in the field of human-tech interaction believe that marketing AI agents as personified ‘employees’ was always a mistake, and created ambiguity about credit and agency. “It inverts our sense of who’s in charge. When an AI tool was framed as an employee, participants in the study saw themselves as less responsible for its output,” wrote James O’Donnell in the MIT Technology Review.
Instead, noted MIT economist Daron Acemoglu, “they should be optimized so that they can improve human capabilities, which is not what they have been at the moment.”
Understanding this dynamic will go a long way to unravel a knot that exists around trust in the world of AI and work, one cited in Canada’s national AI strategy and one likely to be a persistent theme over the next couple years: how do you build trust alongside exploiting the technology?
It is, as yet, an open question, and one that Alessia Artuffo, CEO of education tech firm Docebo, thinks can be a harmful one.
“If employees are producing more output but feeling less ownership of the work, that’s not a win,” he told Business Insider. “That’s capability regression dressed up as efficiency.”
| |
WORKSPACE
Frozen in time
A monument to another era, an abandoned IBM office complex is a strange mix of futuristic architecture and decay — and a reminder of the heyday of the corporate office
| | |
|
ONE OF THE summer’s hottest movies has been the indie hit Backrooms, a psychological movie set in a twisty set of abandoned, supernatural office corridors from the 1990s. Bizarre things ensue, and it does not engender much enthusiasm for returning to the office, that’s for sure. (If you’ve seen the movie, you’ll know what I mean here.)
And yet, while millions of office workers grumble and moan about being told to return to their drab, almost-abandoned office spaces, one particular drab, actually abandoned IBM office from the 1990s in New York state has had no trouble attracting people back. The problem is they are urban explorers, and they keep trespassing to see it.
The site is a famous former IBM campus, which the Wall Street Journal reports “has become a magnet for so-called urban explorers who prowl abandoned malls, hospitals, power plants, amusement parks, factories and any other disused structure they can breach.”
What’s attracting the explorers is not just the challenge of getting in, but the distinct office environment that was left behind: “Eerie glimpses of conference rooms and cubicles unchanged since IBM left, as if employees had fled the daily grind one day and never returned.”
Indeed, like Backrooms, the site’s ghostlike qualities are what makes it most appealing. A decade of slick, modern open-office design naturally creates a desire for its opposite: an architecturally-rendered monument to the heyday of the corporate office, made all the more interesting by the vestigial office stuff left behind — filing cabinets, cubical dividers, staplers and office chairs among them. You might go so far as to say that while workers revolt against the offices they’re being ordered back to, the appeal of this specific IBM campus might prove that the office itself holds at least some deep, psychological pull.
Either way, police (and lawyers) seem to have their hands full dealing with the explorers, many of whom are farming a bit of TikTok content out of the whole thing; they’ve reportedly arrested 48 people for trespassing since February, and many neighbours are fed up with all the action, on top of the vandalism.
Not everyone, though. Some are happy to let the explorers explore, and agree that the office itself — a time capsule of the go-go 1990s office — is probably worth seeing. “It was at one time the Acropolis,” said one town resident. “To think of it as a ruin is absurd.”
|
| | | | |
