Crew / Wikimedia Commons (CC0 1.0)
There are two stories about AI and work. In one, it wipes out millions of jobs. In the other, it makes everyone more productive and creates new work.
Neither matches the full Canadian picture.
A job can change without disappearing. AI can automate tasks, assist workers or change hiring and training.
To see what's actually happening, separate three questions: what AI could do, how Canadians use it, and what employment data shows.
Here's the version that fits in your head.
So what happens when AI becomes part of a job?
Most jobs are made up of several tasks.
A payroll administrator enters information, checks calculations, corrects errors and answers employee questions. AI could process records or flag inconsistencies, but unusual problems still require experience, judgment and someone accountable for the decision.
That creates three possibilities: AI can assist a worker, automate certain tasks or lead an employer to redesign a job. A workplace might use it to free staff for complex cases. Another might decide it needs fewer people for routine processing.
Neither outcome can be predicted from a job title alone.
The distinction to remember: exposed doesn't mean replaced
Researchers use AI exposure to estimate how much of an occupation's work could potentially be affected by artificial intelligence.
According to Statistics Canada's July 30, 2026 release, the March 2026 workforce estimates for Canadians aged 15 to 69 fell into three groups:
31.2% worked in highly exposed occupations where AI has greater potential to complement their work.
29.6% worked in highly exposed occupations with less potential for complementarity, where tasks may face greater automation pressure.
39.3% worked in occupations classified as having lower exposure.
These figures measure potential change, not layoffs. Two occupations can have similar exposure and very different outcomes, depending on responsibilities, regulation and how the work is done. Many trades involve physical work that general-purpose AI cannot perform.
Even where automation is technically possible, cost, reliability, privacy and workplace decisions affect whether it happens.
Exposure tells you where work may change. It doesn't tell you whether a job will disappear.
What's actually happening in Canada?
Canadian data tells us more than global job-loss forecasts.
More workers are using AI, but usually for part of their work
In March 2026, Statistics Canada found that 35.9% of Canadian workers aged 15 to 69 had used generative AI at work in the previous year.
Use was much higher where AI could complement existing work: 53.8% of workers in highly exposed, high-complementarity occupations reported using generative AI, compared with 14.2% in low-exposure occupations.
Among those using it, nearly two-thirds did so for some, but not most, tasks. Only 11.6% used it for most or nearly all of their work. Using AI does not mean handing over an entire job.
Businesses are adopting it too. In the second quarter of 2026, Statistics Canada reported that 19.2% of Canadian businesses had used AI to produce goods or deliver services in the previous year, up from 6.1% two years earlier.
Among businesses using AI, 44.4% changed training or staffing practices. Training existing employees was the most common individual change.
Worker and business figures measure different things. Together, they show AI entering workplaces through employees and employers.
The employment picture is more complicated
A January 2026 Statistics Canada study found that employment generally grew across occupational AI-exposure groups between November 2022 and December 2025. It found no clear evidence of persistent, disproportionate declines in jobs most exposed to AI and least likely to be complemented by it.
Not every worker experiences the transition equally.
In August 2026, Bank of Canada researchers found that unemployed people from highly exposed occupations were having greater difficulty finding new work relative to those from less-exposed occupations. They did not find a comparable increase in workers losing or leaving jobs.
The distinction matters: employers can reduce recruitment or leave vacancies unfilled without announcing layoffs.
These are early signals, not proof that AI caused the hiring difficulties. The pandemic, immigration changes and trade disruptions also shaped the labour market.
We can see some concerning patterns. We cannot yet confidently assign all of them to AI.
Who could feel the changes first?
The effects depend on tasks, not just job titles.
Workers in information-heavy roles may see routine data entry, document processing, scheduling or customer-service tasks automated or reorganized.
People starting their careers may face a different challenge. Entry-level jobs often involve routine work that helps new employees gain experience. If those tasks shrink, the route into more advanced work could narrow. Statistics Canada's January 2026 study also found weaker employment growth among younger and less-educated workers across AI-exposure groups, though it did not establish AI as the cause.
Experienced professionals in law, finance, technology and communications may see more changes to research, drafting and analysis than outright job replacement. Their expertise and responsibility for decisions still matter.
No occupation is automatically safe or doomed. Two people with the same title could experience AI differently depending on their duties, employer and how the technology is introduced.
What about people looking for work?
AI is also changing how some employers hire. They may use AI-based tools to screen applications, assess candidates or support recruitment decisions.
Ontario introduced a disclosure requirement on January 1, 2026. Under the provincial guidance, employers with at least 25 employees generally must disclose when AI is used to screen, assess or select applicants in covered publicly advertised job postings. This also applies when a recruitment firm does the screening, although exceptions exist.
Employers do not have to explain how the system works. And disclosure alone does not tell applicants whether an assessment is accurate or fair.
For job seekers, the practical response is to describe relevant experience and skills clearly, rather than guessing how automated screening works.
What can you actually do?
You do not need to predict which occupations will exist in ten years. Understanding your own work is a better starting point.
Look at your tasks. Which involve writing, researching, organizing information or following repeatable steps? Which depend on your judgment, relationships or experience?
Learn the tools in your field. You do not need to become a programmer. Explore approved AI tools, see where they help and learn their limits.
Keep developing your judgment. Check AI-generated work, explain decisions and recognize when a tool should not be used.
Show what you can do. For job seekers, especially those early in their careers, concrete projects and examples of problem-solving can demonstrate value beyond routine tasks.
Protect sensitive information. Do not enter confidential workplace records, personal information or client material into AI tools unless their use is authorized.
The goal is not to compete with AI at every task. It is to understand where it helps, where it changes the work and where your expertise remains essential.
Our read
Canadian evidence does not establish that AI is causing widespread job losses. It does show growing adoption, changing workplace practices and early signs of hiring difficulties in occupations more exposed to the technology.
The quieter risk is a narrower path into work. If organizations automate routine tasks but reduce entry-level hiring, people starting or changing careers may find fewer opportunities to gain experience, even without widespread layoffs.
That raises important questions for employers and governments: who gets trained, who gets hired, and how will newer workers gain the experience more advanced jobs require?
For workers, the useful response is neither panic nor complacency. It is paying attention to which parts of work are changing and developing skills that remain valuable.
AI at work: what to learn first — Practical habits for using AI effectively and safely at work.
The Weekly Brief — Issue 08 — Bank of Canada research on AI exposure, hiring and the early Canadian labour-market signals.
How to check AI answers and sources — Why AI-generated information needs verification before it is used for important decisions.
