April Hicke writes that Canada is building its AI economy on a foundation that excludes some of its workforce.
April Hicke is the founder and CEO of Toast, a tech recruitment platform dedicated to increasing gender diversity in the industry.
Canada has positioned itself as a global leader in artificial intelligence. We have the talent, the research infrastructure, and the institutional ambition. What we don’t yet have is a plan to ensure that the AI economy works for everyone building it.
Here’s the gap that our national AI conversation isn’t talking about enough: the workers most exposed to AI-driven disruption are overwhelmingly women. The people designing those AI systems are overwhelmingly not. That asymmetry is a structural failure. If Canada doesn’t address it now, while the architecture of our AI economy is still being drawn, we won’t just have a fairness problem. We’ll have a competitiveness problem.The numbers are striking. According to the International Labour Organization (ILO)’s most recent research brief, women-dominated occupations are almost twice as likely to be exposed to generative AI as male-dominated ones. Twenty-nine percent of women’s jobs are at risk, compared to 16 percent of men’s.
A separate Anthropic research report tracking labour market impacts reaches a similar conclusion, finding that workers in the most exposed professions are disproportionately female. At the highest automation risk levels, the gap is even starker: 16 percent of female-dominated roles fall into the most exposed categories, against just three percent of men’s. In high-income countries, Canada among them, the ILO finds that 9.6 percent of women are in the highest AI exposure category, versus 3.5 percent of men. Women in Canada are nearly three times more at risk from automation than men.
RELATED: The compound impact of AI on the labour market
The reasons are structural. Women are heavily concentrated in clerical, administrative, and business support roles: secretaries, receptionists, payroll clerks, accounting assistants. These are roles where tasks are routine, codifiable, and squarely in AI’s crosshairs. Men dominate construction, manufacturing, and manual trades, where physical tasks remain far harder to automate. Meanwhile, women make up just 22 percent of the global AI workforce, and fewer than two percent of venture capital funding in AI-driven startups goes to women-led ventures.
The consequences of that absence show up in the products. Amazon scrapped its internal recruiting AI after it was found to systematically downgrade women’s resumes, having been trained on a decade of male-dominated hiring decisions. Medical AI tools have been shown to perform less accurately for women because the training data skewed male. When the people building these systems don’t reflect the population using them, the systems reflect that absence. The bias is architectural.
That architectural bias does not stay contained to the products being built. It shapes who gets trained, who gets access, and who gets left behind. In Canada, the gap is measurable and growing. A 2024 survey of more than 5,800 Canadians by the Future Skills Centre found that 53 percent of men described themselves as somewhat or very familiar with AI tools used in the workplace, compared to 47 percent of women. More men than women reported receiving AI guidance from their employers, meaning the training gap is not just a pipeline problem—it is happening inside organizations right now, to women already in the workforce.
When employers invest in AI literacy unevenly, they are not just reflecting existing inequity, they are compounding it. The women most at risk of being displaced by AI are also the least likely to be equipped to work alongside it. Canada is simultaneously one of the top AI research and development ecosystems in the world and a country where fewer than four percent of firms have adopted AI in their operations. We are funding acceleration without equity. We are building a talent pipeline without asking who gets to be in it.
Canada cannot build a competitive AI economy on half its talent. Women-owned businesses contribute $150 billion to Canada’s GDP and employ 1.5 million people. Without deliberate intervention, AI becomes a multiplier of the gender pay gap.
Representation needs to be more than a checkbox
There are three things Canada can do now. These recommendations apply with particular urgency to women facing compounding barriers, including those related to race, disability, and immigration status, where the data gaps are wider and the harms of inaction run deeper.
First, mandate gender-disaggregated impact assessments before deploying AI tools that affect workforce decisions such as hiring, performance scoring, promotion, and compensation. If you cannot show how your AI system performs across gender, you should not be deploying it at scale.
Second, fund women’s access to AI upskilling at the same rate and with the same urgency as AI research and development investment. Every dollar Canada invests in AI acceleration should have a corresponding commitment to equitable access to the skills those investments create.
Third, require women’s representation as a condition, not a checkbox, in AI governance, ethics boards, and federal AI advisory bodies. The decisions being made in those rooms will shape the economy we all, including the women typically excluded, will inherit. They should be included in making those choices.
The ILO notes that the impact of AI on women’s jobs is not predetermined. Canada has built something rare: a world-class AI ecosystem with the values and institutional capacity to lead responsibly. The architecture of our AI economy is still being designed. The question is whether we’ll design it for everyone.
The opinions and analysis expressed in the above article are those of its author, and do not necessarily reflect the position of BetaKit or its editorial staff. It has been edited for clarity, length, and style.
Featured image courtesy Unsplash. Photo by Christina @ wocintechchat.com.
