The federal government is aiming to increase the use of AI among Canadian businesses and in the Canadian economy by 500 percent over the next five years. If it accomplishes this goal, it will be an unprecedented technological and systemic invasion of Canadian workplaces. Workers are distrustful and wary, and labour unions are playing catch up to be informed, understaxnd and cope with the impacts.
No need to worry, Minister Evan Solomon is telling Canadian workers. Canada’s AI for All policy assures us that the build-out of AI “will be powered by workers” and “over 250,000 new AI-relevant jobs will be created across the Canadian economy by 2031.”
Almost no one believes that. In May, an Angus Reid survey of white collar, blue collar and retired Canadians, found that only three per cent believed that AI will significantly increase jobs. A 45 per cent plurality of those surveyed expected a significant loss of jobs.
Not even the proponents of big data centres are claiming that the AI build out will result in significant new direct employment. The Bell AI Fabric proposal for an AI data centre in Sherwood, Saskatchewan, claims it will be the largest AI facility in Canada. But it isn’t trying to promote the project based on its underwhelming employment projections: 800 temporary construction jobs and 80 ongoing operational jobs, plus several hundred unspecified indirect jobs in the community.The somewhat smaller Telus AI proposal for two data centres in Vancouver, and an expansion of an existing facility in Kamloops, likewise touts more than 1,000 construction jobs and “hundreds” of permanent operating jobs.
Of course, there is a fundamental reason for the dissonance between Solomon’s optimistic projections and everyone else. The purpose of AI technology is to increase productivity by aggregating knowledge, performing unprecedented large-scale data tasks, and greatly increasing labour productivity. That can mean a larger economy in the long run, but also less workers when AI is deployed in a workplace or industry.
Impacts on the labour market so far
There have not yet been large-scale job losses in Canada resulting from AI. In 2024, six per cent of Canadian firms that have implemented AI reduced employment as a consequence. By mid 2025, the number of AI adopting firms that expect job losses had jumped to 12 per cent. These are still relatively small numbers, but so too is the take-up of AI systems by Canadian businesses—less than 15 per cent of all businesses in 2025.
However, a KPMG survey of Canada’s biggest corporations reports a whopping 93 percent have begun investing in and implementing AI. They are just getting started, though; only 31 percent of those companies have fully embedded generative AI into their operations, and 98 percent have not yet had net returns from their AI investments. That will change as business moves to “agentic AI” that operates independently and drives “operational efficiency.”
At least one outlier report predicts much larger consequences. The Signal 49 report (formerly the Conference Board of Canada) estimated in January 2026 that AI automation could leave employment as much as 555,000 jobs below what it otherwise would have been by 2030, with a sustained recovery taking almost a decade before the productivity gains of AI eventually exceed projections without the AI build-out. To put that in perspective, Canada lost about 425,000 jobs in 2008-2009 after the financial crisis. A BNN Bloomberg story on the government’s response to the Signal report was not reassuring: “In a technical briefing with reporters …, government officials said they don’t contest or agree with the Conference Board of Canada report but will monitor the impacts of potential displacements.”
Signal 49 also projected that the impact of AI will be felt across the economy. Although the services and sales sector is the least AI exposed, it is the largest employment sector still affecting large numbers of workers. Other employment sectors like applied sciences, manufacturing, trades, transport, equipment operators, and agriculture are more AI exposed. Health care, education, business and finance land in the middle. The result is that the broad sweep of the Canadian working class will have a common struggle for job security and just transitions.
Is the Signal report overstated? Maybe. In his new book, The Reverse Centaur’s Guide to Life After AI, Canadian writer and pro-labour technology analyst Cory Doctorow situates the fight over AI in the workplace and over who controls its use. He argues against inevitable consequences. Don’t fight AI, he advises, “fight the boss.”
Doctorow contends that AI companies and corporations deliberately exaggerate AI’s ability to displace workers to attract the large investments needed to fuel the AI build-out. The gargantuan investments by AI companies require continuous waves of new investments, and that means convincing Wall Street investors that AI will “replace vast swathes of the wage-earning human workforce.”
A growing consensus is that the largest economic threat that AI poses is the popping of the investment bubble currently driving the U.S. economy. “When the AI bubble pops, it will cause a recession and major job losses in the U.S.. Overheated capital spending on data centres and power plants, and excessive luxury consumption by those who’ve been made rich (on paper) by the speculative flight of the market, will quickly shift into reverse. That downturn will spill over into Canada,” Centre for Future Work Director Jim Stanford warns.
Legislation and collective agreements
Regardless of the actual extent of AI worker displacement and net employment effects, AI has opened a new frontier for worker rights. The 2026 Canadian Labour Congress convention resolution on AI revealed the breadth of issues facing workers and their unions, and the extent of work to be done. The composite resolution, combining seven separate resolutions, called for establishing a CLC committee to review bargaining and arbitrations on AI, developing an AI policy to protect jobs, and addressing a litany of concerns ranging from training and transparency to fair compensation, surveillance and algorithmic management, human rights protections, regulatory oversight, and liability for technology companies.
Unions are already having to put AI on the bargaining table. In telecommunications, despite collective agreement language to prevent layoffs resulting from technological change, Telus has cut 11,000 Canadian jobs and Bell about 8,000 since 2023. The telecoms will not attribute the job losses to technological change and instead point to changing customer service models and operational efficiencies. The telco unions know exactly what is going on. In the latest clerical agreement with Bell, Unifor negotiated a Joint Committee on Artificial Intelligence to review the introduction of AI in the workplace, and stronger technological change provisions.
However the only practical response to a possible AI bomb dropping on Canadian workers in the short to medium term will be legislative. Canada’s three telecom unions—Unifor, USW and CUPE—told a parliamentary committee on AI regulation in June that a “comprehensive social dialogue” and a permanent sectoral working group to regulate the use of AI is needed.
In the case of the Canada Labour Code, reform is long overdue. Sections 51-55 of the code on technological change were written in 1973 and have not been substantially amended since. You can drive a fleet of Uber Teslas through the gaps that have arisen. The technological change provisions in Canadian labour law do not capture the scope of AI, and in any event merely oblige the employer to give notice, detail expected consequences and provide a rationale for the introduction of the new technology.
Some forms of AI in Canadian workplaces will meet even the very limited legal and collective agreement definitions of technological change. However, AI is more than a robot on an assembly line, or a new machine or accounting system that needs half as many workers. AI can potentially displace entire functions that employers deem unnecessary. The employer may claim that the new AI system is a new function never before carried out by the displaced workers. Not only workers, but whole departments and the managers they formerly reported to could be replaced. Of course these are all issues of bargaining and adjudication, but the scales of justice weigh heavily towards employers, and the absence of AI regulation or modern labour laws add many thumbs to the scale.
For the majority of the working class which has no collective agreement coverage, and hence is protected only by employment standards legislation, there are no laws on technological change and only the beginnings of recognizing a new world of AI-driven work.
Recent employment law changes in BC recognize a new category of AI driven “platform work” to ensure minimum wages and workers’ compensation coverage for platform workers, as well as pay transparency provisions to address algorithm-based pay. Ontario has enacted employment standards requiring disclosure if AI is used to screen, assess or select workers in hiring and employment matters, and its new Digital Platform Workers’ Rights Act requires minimum wage and pay transparency for platform workers. But the Ontario gig worker wage provisions are not real minimum wages, and the Ontario Act does not recognize platform workers as employees like the BC Act.
All these nascent regulations are new, incomplete and far from becoming national standards. At the federal level there are no worker rights laws or standards on AI. The AI for All strategy proposes no constraints or even monitoring of employment consequences, or any intention by the government to require disclosure, consultation, or negotiation regarding AI in the workplace.
Those omissions are glaring when, only weeks before, the government’s consultation on the Canada Labour Code included the need for AI regulations as a question among a lengthy list of major issues from the right to strike to sectoral bargaining. However, all of that was compressed into a 30-day consultation, widely panned as insincere and a cover for the government’s intention to restrict strikes at ports, airlines and railroads.
The battles ahead
The AI invasion should be an urgent catalyst for labour to assemble an army to march on Ottawa and claim the influential role it deserves in AI regulation and labour law. It will need to be more than a polite lobby and come armed with extensive briefs and demands on behalf of all Canadian workers who are about to feel the invisible hand of a new and unregulated economic force.
But barring a startling political shift to slow and alter the AI trajectory, worker outcomes on AI are more likely to be settled piecemeal in multiple workplace, industrial and sectoral battlegrounds.
The seminal labour battle to date over AI is undoubtedly the lengthy and hard-fought strikes across the U.S. by the Writers Guild of America and the Screen Actors Guild in 2023. The WGA agreement that resulted makes AI a tool controlled by the worker—not a mandatory management technology. The agreement affirms that AI is not a writer and cannot do the work of a writer, nor can they force a writer to use AI. The SAG agreement requires consent and compensation for digital replications of actors, and consent for being photographed or scanned for that purpose. All performers, including background actors, must consent and be compensated for any digital replication and any future use of their replication, including after their death.
The writers’ and actors’ strike underscored the power of solidarity to confront generational challenges before larger political and legislative norms are established. The terms of those agreements also tell us that AI is not just a new kind of technological change. The nature of work and the human integrity of workers are at issue.
It will take decades of collective bargaining and pivotal labour campaigns and strikes to determine the terms and conditions of AI in the Canadian world of work. This work begins with tech companies and employers far ahead. To catch up, labour must first imbue workers with a deep understanding of how AI affects their work and then forge a class-based bargaining strategy and political program for AI regulation and labour law reform.
About the author
Fred Wilson
Fred Wilson is retired Director of Strategic Planning with Unifor (previously Assistant to the President with the Communications Energy, and Paperworkers Union of Canada). Author of A New Kind of Union (Lorimer, 2019). Follow his writing on labour and issues at medium.com/@jefred.wilson
