What to know
- The findings describe reported use, not measured job displacement.
- A task count does not tell us how much working time a tool changes.
- For career decisions, ask which responsibilities change and who remains accountable.
What the survey found
IVADO's October 15 announcement highlights a CRIMT study in which 43% of respondents reported using generative AI at work. The task-share figure covers users only; that chart lists 494 respondents. Percentages are weighted.
Who was surveyed
The online survey covered 1,015 Quebec workers from the nonprobability Léger Opinion panel, June 26–July 2, 2026. Researchers weighted responses by gender, region, age and education. This is self-reported, cross-sectional evidence, not a causal employment study or a representative global estimate.
Sources: CRIMT
Our interpretation: look beyond the adoption label
For someone comparing careers or evaluating a workplace, a yes-or-no adoption label leaves the important question unanswered: what has actually changed? Generating a first draft, checking a calculation and deciding whether to approve a sensitive request involve different responsibilities. Counting each as one task would not capture the time, stakes or expertise involved.
Our practical reading is to ask employers for a concrete account of the work. Which steps now use a tool? Who checks the result, handles exceptions and explains the decision? A role can retain its title while those responsibilities move between colleagues. Conversely, adding a tool to one routine step need not transform the rest of the job.
For managers, this suggests a useful reporting distinction: access to a tool, actual use and a demonstrated improvement are separate milestones. Track them separately before changing a workload target. For workers, describing the judgment and verification behind an AI-assisted result can make a skills discussion more specific than simply listing a product on a résumé. These are editorial suggestions, not effects tested by this survey.
Sources & method
AI-generated article checked against the release and full report. The report is dated October only; October 15 dates the announcement. Its separate Reddit analysis is not used here. Practical implications are original editorial interpretation.
- IVADO announcement of the worker study IVADO ·
- Worker perceptions and professional impacts: full French report CRIMT ·
Drafted and source-checked by Codex AI agents. Read our sourcing, illustration, and correction policy.
Explore the occupations
What could this mean for your work?
See task exposure and potential career moves for roles connected to this story.
