How many Canadian workers are highly exposed to AI, and where?
We combined what public data exists: Anthropic's measurements of which tasks AI actually performs, Statistics Canada's worker and business surveys, and job-posting data. This page states what the evidence shows, what it cannot show, and which trades and skills the data points to.
What Anthropic actually measured
Anthropic's Economic Index mapped millions of anonymized Claude conversations onto the U.S. Labor Department's O*NET task statements. That makes it a measure of observed use, not expert opinion. The March 2026 observed-exposure paper then combined that usage with theoretical task feasibility.
Who is exposed in Canada
StatCan's Mehdi and Morissette (2024) adapted the AI occupational exposure index to Canadian occupations, and split workers by whether AI is likely to augment or compete with them.
Adoption is climbing, unevenly
Business adoption tripled in two years, but it concentrates in information, finance, and professional services. Construction sits at 9.2% and agriculture at 4.5%.
Where exposure concentrates
Indeed Hiring Lab scored Canadian metros by the share of local job-posting skills that generative AI could transform. Treat it as directional: the underlying skill-exposure data is U.S.-based.
The entry-level squeeze
StatCan's Mehdi and Frenette (January 2026) found that in coding-intensive occupations, employment of workers under 30 stagnated from November 2022 to December 2025 while ages 30 to 49 kept growing. The mechanism is straightforward: firms keep the senior who directs the AI and stop hiring the junior whose tasks the AI now handles.
At risk, and likely to thrive
A three-to-five-year view, combining observed task coverage, exposure scores, and Canada's occupational mix.
Most at risk of substitution
- Data entry and office support67.1% observed task coverage, the highest measured clerical exposure.
- Customer service70.1% observed task coverage; chatbots already absorb routine tickets.
- Routine finance, accounting, bookkeepingConcentrated in Toronto's finance sector, which leads business AI adoption at 40.4%.
- Translation and bilingual servicesLanguage tasks are core LLM strengths; Ottawa-Gatineau demand is exposed.
- Junior and entry-level software developmentUnder-30 coding employment stagnated after 2022 while senior hiring continued.
- Graphic design and routine content productionWEF lists it among declining roles for the first time, linked to generative AI.
Most likely to thrive
- Skilled trades: electricians, plumbers, carpenters, welders14.7% GenAI use in trades and transport; physical work stays grounded in the real world.
- Healthcare: nurses, nurse aides, personal support workersAccountability and physical care keep humans in the loop; nurse aides rank among top growing occupations.
- AI-augmented professionalsSurgeons, judges, senior engineers: high exposure, high complementarity, where AI raises the expert's output.
- AI builders and AI-fluent operatorsAI-user roles grew by about 62,000 postings in 2025; the premium goes to people who direct the tools.
- Sales, commercialization, relationship rolesTrust and persuasion resist automation; PwC finds exposed jobs adding empathy and judgement tasks 2.5x faster.
AI jobs are growing too
Canadian postings requiring AI skills rose from about 60,000 in 2024 to 129,000 in 2025, from 1.6% to 3.5% of all postings. AI-user roles, people applying AI inside existing business functions, grew by about 62,000.
What training actually helps
Canadian sources converge on the same curriculum. It is not a coding bootcamp for everyone.
AI fluency inside your current job
Learn the tools on the job, alongside the technology. A TD Bank survey finds on-the-job training alongside the technology is the most effective format.
Judgement and relationship skills
Critical thinking, ethical judgement, relationship management. Institut du Quebec: stronger soft skills produce better prompts and better validation of AI output.
A license or credential where liability stays human
Healthcare, licensed trades, regulated professions. Where accountability cannot be automated, the credential is the moat.
Commercialization skills
Selling, scoping, and delivering work. RBC: demand is growing for coordination, social perceptiveness, and complex problem solving.
The Future Skills Centre's IVADO program trained more than 3,000 workers with job-adapted AI courses, a self-diagnosis tool, and certification. The Centre's summary line: workers are replaced by other workers with better AI skills, not by AI alone. Its broader guidance: basic AI skills should be core digital literacy for every field, including knowing when not to use AI. The Labour Market Information Council adds adaptive human-in-the-loop skills for collaborating with AI, and reports that Canadian companies are rapidly adopting AI while most employees are not getting trained to use it. Policy Magazine's warning: Canada's workforce AI training is fragmented and not standardized, and SMEs lag behind.
Methodology and limits
What this combines
- Anthropic Economic Index usage data mapped to O*NET tasks: observed use across millions of conversations.
- StatCan business surveys (CSBC, 2024 to 2026) and the July 2026 worker generative-AI survey.
- StatCan analytical papers: Mehdi and Morissette (2024), Mehdi and Frenette (2026), Leanage and Mehdi (2026).
- Job-posting data from Indeed Hiring Lab and PwC's AI Jobs Barometer.
What is still missing
- No Canada cut of Anthropic's Economic Index has ever been published.
- No public crosswalk links Canada's NOC occupation codes to O*NET task data.
- The Labour Force Survey does not ask workers about AI use on the job.
- AI-talent emigration is not measured at the occupation level.
Sources
- Anthropic, "Which Economic Tasks are Performed with AI? Evidence from Millions of Claude Conversations." January 2025.
- Anthropic, task-level collaboration patterns (Claude 3.7 Sonnet window). March 2025.
- Anthropic, Massenkoff and McCrory, observed job-exposure paper. March 5, 2026.
- Anthropic, "Scenarios for our Economic Future." September 9, 2026.
- Statistics Canada, Canadian Survey on Business Conditions AI modules (tables 33-10-0825-01, 33-10-1004-01, 33-10-1167-01). Q2 2024, Q2 2025, Q2 2026.
- Statistics Canada, worker-level generative AI use survey. July 30, 2026.
- Mehdi and Morissette, "Experimental Estimates of Potential Artificial Intelligence Occupational Exposure in Canada," StatCan Economic and Social Reports. 2024.
- Mehdi and Frenette, "Canadian employment trends in the era of generative artificial intelligence: Early evidence," StatCan. January 28, 2026.
- Leanage and Mehdi, "Potential occupational exposure to artificial intelligence and automation among certified journeypersons in Canada," StatCan. January 2026.
- Indeed Hiring Lab, Canadian metro AI-exposure analysis. October 2026.
- PwC, AI Jobs Barometer, Canada cut. 2026.
- Future Skills Centre / Conference Board of Canada, "Understanding the Influence of AI on Employment." January 2026.
- World Economic Forum, Future of Jobs Report 2025. January 2025.
- IMF, "Gen-AI: Artificial Intelligence and the Future of Work," Staff Discussion Note 2024/001. January 2024.
- Kickresume, AI job-replacement risk analysis of 756 jobs. September 2026. MDPI, 2026 paper on electrician task impact by 2038. U.S. Bureau of Labor Statistics, Occupational Outlook Handbook, electricians. 2025 to 2035 projections.
- RBC, "Humans Wanted: How Canadian Youth Can Thrive in the Age of Disruption." March 2018. TD Bank workforce AI training survey; Institut du Québec on soft skills and AI use; Future Skills Centre, "Unleashing AI into the Skills Development Ecosystem"; LMIC on the credential boom and the adoption-training gap. 2026. Policy Magazine on fragmented AI training in Canada.