The Federal AI Strategy Rollout
OTTAWA — The Canadian government’s long-promised AI strategy landed Thursday. It’s called “AI for All.” The 50-page document promises more than $2 billion in new spending and sets a target of 60 per cent of businesses using AI within a decade. The parts that would actually shield people from deepfakes, job loss, and chatbot-assisted violence are barely there.
At a Toronto hospital launch, Prime Minister Mark Carney told the crowd, “AI could be weaponised against us. We have to be honest about the risk that AI poses to Canadians.” That blunt talk doesn’t show up in the plan, opposition critics say. They’re not wrong.
Billions in spending, big job promises
The government’s commitment is $2 billion (roughly $2.8 billion CAD — briefing materials also cited a US$1.4 billion equivalent, which doesn’t simplify things). Half a billion of that would go to direct investments in Canadian AI companies. The government would take equity stakes in some. Another $500 million in financing will push businesses to adopt AI, and $50 million is set aside for creators so they can use the tech “on their own terms.”
The strategy vows to create 250,000 jobs by scaling AI across sectors. But there is no estimate anywhere of how many jobs might vanish. Union leaders and economists flagged that gap almost immediately.
The adoption target is steep. According to Statistics Canada data cited in the strategy, 12 per cent of businesses were using AI between mid-2024 and mid-2025. The plan wants that number to hit 60 per cent by 2034.
Sovereignty, a supercomputer, and the brain drain
The document concedes that Canadian data often sits on foreign servers and that the government itself relies on infrastructure it doesn’t control. To fix that, the plan proposes a “world-leading” public supercomputer and large-scale AI data centres on Canadian soil, with capacity ramping up significantly by 2030.
It also calls the brain drain to the U.S. an “uncomfortable reality.” It names names: Geoffrey Hinton, the Nobel-winning “godfather of AI,” sold his firm to Google. Ilya Sutskever, co-founder of OpenAI, crossed the border. The strategy offers more fellowships, more university research chairs, and a fast-track immigration path for AI specialists to try to reverse the flow.
Health care and a trip to the library
The plan puts $200 million toward AI in health care. That’s tricky federal ground, since health is mainly a provincial responsibility, but the government points to Europe: Carney noted that Medical imaging is where AI has made its deepest inroads across Europe — close to three in four EU member states have folded AI diagnostic tools into that work.. The hope is that AI can drive down doctors’ administrative workload.
Then there’s the public trust problem. Government polling shows half of Canadians see AI as a threat to humanity, and 36 per cent consider it harmful to society. So Ottawa wants to partner with public libraries to offer “entry-level AI training.” Canada ranks low among peer countries on AI literacy and trust. The strategy admits that far.
Safety? Details are missing
Conservative MP Melissa Lantsman summarized the opposition’s complaint: “The safety and security that was promised in this is nowhere to be found in the document, certainly no details.” The government says it will bring in new AI laws to protect consumer privacy, children, and to update online safety legislation. That’s it. No timeline, no draft bill, no indication of what enforcement powers a regulator would have.
The gap isn’t theoretical. Earlier this year, a suspect in a mass shooting in Tumbler Ridge, B.C., used ChatGPT to discuss gun violence months before the attack. OpenAI knew and didn’t alert police. The CEO later apologized, and company executives were summoned to Ottawa. That episode shook public confidence, and it makes the absence of safety provisions in the new strategy all the more glaring.
One digital rights researcher, who wasn’t authorized to speak publicly, put it bluntly: “You can’t promise a revolution in AI adoption and then admit that deepfakes, unsafe chatbots, and disinformation are proliferating without telling Canadians exactly what safeguards will be in place by what date.” A separate internal government plan for how the public service will use AI—released the same week—covers central capacity, governance, talent, and transparency, along with a new Centre of Expertise. But that’s about the bureaucracy’s own operations, not the broader societal guardrails.
What comes next
Carney’s office says legislative proposals will land in the House of Commons this autumn. No guarantee they pass before an election, though. The industry committee gets a briefing in two weeks.
Creative destruction happens whether governments publish strategies or not. For now, Canadian AI researchers and venture capitalists are waiting to see if the investment and immigration promises move fast enough to stem the talent drain. And for the millions of Canadians who’ve never touched an AI tool, the most concrete thing in this 50-page plan might just be the bit about training at the local library. That isn’t exactly a firebreak against deepfakes.
