The Weekly Brief — Issue 08

Hello,

For most of the AI-and-jobs debate, Canadians have been asked to think about what might happen.

This week, the Bank of Canada gave us some of the first evidence about what may already be happening.

Its researchers found that people coming from occupations more exposed to AI appear to be having greater difficulty finding work than they did before 2019.

What they did not find was a comparable increase in people losing their jobs.

That distinction matters.

Plus: Anthropic is staffing up for gigawatt-scale computing infrastructure in Canada, a First Nations-led health hub in Saskatchewan gets another $5.4 million, and OpenAI is rolling out a dedicated ChatGPT experience for teens.

here is the week.

What happened

Bank of Canada researchers published new analysis examining whether occupations more exposed to artificial intelligence are beginning to experience different labour-market outcomes in Canada.

They combined an AI-exposure measure adapted to Canadian occupations with individual-level data from Statistics Canada’s Labour Force Survey.

The result is cautious, but important.

Canada has not experienced a broad AI-driven restructuring of its labour market. But people looking for work after leaving occupations with greater exposure to AI appear to be finding it increasingly difficult to get hired.

Using an upper-bound regression comparison between hypothetical occupations at 100% and 0% AI exposure, the Bank estimates that the job-finding gap widened from 2.2 percentage points in 2015–19 to 13.9 percentage points in 2025.

Over the same period, the estimated difference in job-separation rates remained essentially unchanged.

Why it matters

The most common question about AI and work is still: Will AI take my job?

The Canadian evidence suggests that may be too narrow a question.

Companies do not need to eliminate large numbers of existing positions for AI to change the labour market. They can hire fewer people into some roles, combine responsibilities, change the skills they expect from applicants or use AI to handle parts of work that previously required additional staff.

Those changes are quieter than layoffs, but they can still reshape who gets hired and what employers value.

A second Bank of Canada analysis adds the employer side of the picture. Only 8% of surveyed businesses said they use AI significantly in their core operations, suggesting the transition is still early. But looking three years ahead, 23% expect AI to reduce employment, compared with 11% expecting a positive employment effect.

The Bank also found that roughly 30% of businesses expect AI to increase capital spending over the next year, rising to almost 40% over three years. In other words, many firms appear to be preparing to invest in AI before they expect large changes in employment.

The early signal is showing up much more clearly in hiring than layoffs.

The Bank also found that younger workers are disproportionately represented in some occupations with moderate or high AI exposure, including customer-service and sales-support roles.

That does not mean AI caused Canada’s youth-employment challenges. But it gives younger workers, employers and policymakers another reason to pay attention to how entry-level work is changing.

Your move

Do not make a career decision because of one study.

Instead, look one level below your job title.

Write down the five or six tasks that make up most of your work. Which are routine information-processing tasks that AI can increasingly assist with? Which depend on judgment, relationships, accountability, specialized expertise or physical work?

Then ask a better question than “Can AI do my job?”

Ask:

“If AI changes part of my job, what becomes more valuable for me to be good at?”

That is the transition worth preparing for.

Anthropic is planning for gigawatt-scale computing in Canada

Anthropic is recruiting a Compute Country Lead for Canada to oversee its Canadian infrastructure strategy—from site selection and energy through financing, construction and operations.

The company says the role will ultimately be responsible for bringing gigawatts of computing capacity online in Canada, with a footprint intended to scale from initial sites toward multiple gigawatts.

A separate Anthropic position is based in Alberta and focuses on community engagement in places where its data centres are sited, built and operated, including issues around power, water and land use.

The verified materials reviewed for this issue do not identify a final Canadian site, investment amount or construction schedule.

Why it matters: Canada has spent heavily trying to expand domestic AI computing capacity. Anthropic’s hiring provides direct evidence that one of the world’s major AI developers is planning a substantial Canadian footprint of its own.

The next questions are where it builds, how those sites are powered and what governments require—or offer—to bring them here.

A First Nations-led health hub gets another $5.4 million

Ottawa announced $5.4 million in additional funding for the Virtual Health Hub in Whitecap Dakota Nation, Saskatchewan.

The money will support its 2026–27 operating budget and comes on top of more than $28.2 million Indigenous Services Canada says it has already invested toward construction.

Expected to open in early 2027, the First Nations-led facility plans to use telehealth, artificial intelligence and robotics to connect people in rural, remote and northern communities with health professionals and specialists.

The federal government says the hub is expected to support 80 jobs and initially have capacity to serve 30 communities, potentially expanding to as many as 90 by 2029.

Those remain projected outcomes; the hub is not yet operational.

Why it matters: This is AI being applied to a very Canadian public-service problem: how to make specialized health care more accessible when patients and clinicians are separated by enormous distances.

Just as important, the model is First Nations-led and includes training intended to build Indigenous health-care capacity alongside the technology.

ChatGPT is rolling out a dedicated experience for teens

OpenAI launched ChatGPT for Teens this week.

Users who say they are between 13 and 17—or whom OpenAI’s age-estimation systems identify as likely under 18—are placed into a dedicated teen experience.

OpenAI says it includes stronger default protections, parental controls, Study Mode, break prompts and additional restrictions around higher-risk content.

ChatGPT for Teens is rolling out globally to eligible accounts, including in Canada.

Why it matters: Young people are already using general-purpose AI tools. The increasingly important question is not simply whether they use them, but what safeguards, education and adult oversight surround that use.

The new experience is a product-design change, not proof that the underlying risks have been solved. How well the protections and age-estimation systems work in practice will matter more than the feature list.

AI exposure is not the same thing as automation—or job loss.

Researchers look at the tasks people perform within an occupation and estimate how many of those tasks current AI systems could meaningfully affect.

A job heavy on information processing—organizing, summarizing, drafting, searching or transforming information—may therefore have relatively high AI exposure.

But a highly exposed occupation can still require people.

AI might automate one task, speed up another and leave several others largely unchanged.

That is why exposure scores should not be read as forecasts of how many jobs will disappear.

They are better understood as a measure of how much the work itself may change.

Choose one ordinary workday and write down the five tasks that take the most time.

Beside each one, mark:

Then choose one task in the first category and test whether AI can make the work better without removing your review.

The goal is not to automate everything.

It is to understand where AI actually changes your work—and where it does not.

A calm, practical briefing on the AI developments that matter in Canada—delivered every Tuesday.

That’s the week. See you next Tuesday.

— Padge T.

AI Brief Canada is an editorially independent publication published by Groundshift Advisory Inc. and written by Padge T.

Groundshift Advisory helps Canadian organizations adopt AI responsibly. Sponsored, affiliated, or commercially connected content will always be clearly disclosed.

Issue disclosure: Nothing in this issue is sponsored, affiliated, paid, or commercially connected.

AI use: AI tools supported research, source discovery, outlining and drafting for this issue. All material was reviewed, edited and fact-checked by Padge T. before publication. Read our AI Use Disclosure.

Not affiliated with the Government of Canada.

Corrections or feedback: We welcome corrections and reader feedback. Email [email protected] or review our Editorial Policy.