Carney National AI Strategy to Launch Next Week With Six Pillars

Prime Minister Mark Carney confirmed on Wednesday that the federal government's long-delayed national artificial intelligence strategy will be released within the next week, ending months of speculation about how Ottawa intends to position Canada in a global race that has been moving faster than most governments can legislate. Speaking to reporters, Carney described the framework as a balanced document built around six pillars, signalling a clear shift from the adoption-first messaging the Liberals used earlier in the year toward a strategy that weighs safety, sovereignty, and labour disruption alongside economic opportunity.

The new strategy is expected to land just days before the start of the federal summer recess, giving the government a chance to set the tone of the AI debate before parliamentarians scatter for constituency work. AI Minister Evan Solomon, who has been the file's public face, has said in recent weeks that Ottawa wants to strike a balance between AI enthusiasts and skeptics, a tone that has been mirrored in the Prime Minister's own remarks. Carney told reporters that the strategy will be a living document, and that supporting regulation and funding announcements will follow through the summer and fall.

The six pillars Carney outlined

According to the Prime Minister's office, the strategy will be organised around six themes. Protecting Canadians and safeguarding democracy comes first, a framing that gestures directly at concerns about deepfakes, election interference, and online safety. Powering AI adoption for shared prosperity is the second pillar, capturing the productivity case that Carney has personally championed since his days running the Bank of England.

The other pillars focus on building what officials describe as a sovereign AI foundation, scaling Canadian champions, and building trusted partnerships and global alliances. Government sources have described the sovereignty pillar as encompassing data centres, compute capacity, and domestic models, an acknowledgement that Canada cannot continue to rely entirely on American hyperscalers for the underlying infrastructure of its digital economy. The partnerships pillar is expected to outline how Canada will work with allies like the United Kingdom, the European Union, and Japan on shared safety standards.

A sixth pillar, focused on workforce transition and skills, is expected to land alongside a renewed labour market component. Solomon has flagged that the strategy will consider the technology's impacts on workers, a politically sensitive issue given that early adopters of generative AI have already begun pruning entry-level white-collar jobs in marketing, customer service, and software development.

Why the strategy has taken so long

The previous Liberal government floated an AI strategy refresh in 2023, but a federal election, a change of prime minister, and a sharp tonal shift in the public conversation about AI safety have all pushed the timeline back. An advisory group struck under the previous administration was widely criticised for tilting too heavily toward industry voices and not enough toward academics, civil society, or labour. Several rounds of consultations followed, and Solomon was eventually given the file as part of Carney's cabinet reshuffle after the April 2026 majority win.

In the meantime, the United States, the European Union, the United Kingdom, and a handful of middle powers have all moved on AI legislation or executive frameworks. The European Union's AI Act is now into its first compliance deadlines, the United Kingdom has stood up a permanent AI Safety Institute, and Washington has produced a series of executive orders and procurement standards that have effectively set the floor for major suppliers. Canadian companies seeking to do business in those markets are already navigating overlapping rules.

Industry groups have been pressing Ottawa to provide clarity, in particular on liability for foundation models, copyright, and the rules around using Canadian data for training. The strategy is not expected to resolve all of those questions, but it is expected to set out a sequencing roadmap and identify which pieces will require new legislation versus regulation under existing frameworks.

The political backdrop

Carney is unveiling the strategy at a moment when his government is enjoying its honeymoon majority but also navigating intense pressure on multiple fronts. The Trump administration's tariffs on Canadian steel, aluminum, and softwood lumber continue to bite, the Permanent Joint Board on Defence has been suspended, and provincial premiers are demanding a tighter federal-provincial framework on everything from energy to housing. Folding AI into a broader sovereignty narrative is consistent with how Carney has been talking about energy and industrial policy since taking office.

The Prime Minister has made plain that he sees AI as central to closing Canada's chronic productivity gap with the United States. Earlier this month, he linked the strategy explicitly to Ottawa's effort to fast-track major projects, framing AI infrastructure as the same kind of nation-building investment as pipelines and transmission lines. Critics on the left have warned that the framing risks letting industry write its own rules; critics on the right have argued that the government is already moving too slowly relative to private sector deployment.

Reaction from opposition and industry

Conservative AI critic Michelle Rempel Garner said in a statement that the Liberals had spent two years promising a strategy and that the test will be whether the document delivers concrete investments and timelines rather than aspirational language. The NDP welcomed the focus on safety and democratic protection but warned that workforce transition cannot be left to general assurances, calling for a dedicated displaced-workers fund tied to AI adoption.

Industry reaction has been more measured. The Council of Canadian Innovators, which represents domestic technology scale-ups, said in a brief statement that the strategy's success will hinge on whether procurement and immigration rules are aligned with the sovereignty pillar. Several Canadian foundation model developers and chip-design firms have signalled in recent months that they would welcome a clearer Buy Canadian preference in federal AI procurement, particularly for sensitive workloads in defence, health, and intelligence.

The compute and research ecosystem

One of the most concrete elements of the strategy is expected to be the expansion of public purpose compute capacity. Canada's existing public compute infrastructure, while well regarded among researchers, has been overwhelmed by the demand generated by the rise of large language models and other foundation model architectures. Federal officials have indicated that the strategy will commit to a multi year expansion of the Compute Canada successor system, with new facilities planned for several provinces.

The strategy is also expected to expand the Canadian AI Safety Institute, which was established under the previous government and has been operating with a modest budget. Federal officials have indicated that the Institute will be given new statutory authority and additional funding to participate more fully in international AI safety standards work. The Vector Institute in Toronto, MILA in Montreal, and Amii in Edmonton are all expected to receive renewed federal support as part of the package.

What it means for Canadians

For most Canadians, the strategy will not produce immediate visible changes, but it is likely to shape the technology layer of daily life within a couple of years. The democratic protection pillar will probably accelerate work on watermarking and disclosure standards for AI-generated political content ahead of any future election, including the next Ontario provincial vote and the Quebec general election scheduled for October.

The adoption pillar is expected to drive AI deployment across federal departments, including immigration processing, tax administration, and health data analysis. The Prime Minister's office has been careful to emphasise that automation in government will be paired with human oversight, in part because earlier failed deployments such as the Phoenix pay system have made the public service institutionally cautious. The workforce pillar will eventually feed into a new federal training program intended to help mid-career workers retool.

The sovereignty question

The most distinctive element of the strategy may be how aggressively it leans into AI sovereignty. Government officials have hinted in background briefings that the document will include language on a domestic compute capacity build-out, modelled in part on the Canadian Sovereign AI Compute Strategy launched in 2024 but with much larger ambitions. Ottawa is expected to commit to a long-term funding envelope for public-purpose compute that universities, researchers, and small Canadian firms can access without depending on American cloud providers.

Canada is also expected to use its G7 presidency in 2025 and its remaining diplomatic capital to push for interoperable safety standards. Officials have described that effort as the Canadian version of the small-country playbook used effectively on climate and tobacco regulation, where Ottawa punches above its weight by convening like-minded middle powers.

International alignment

The strategy's international pillar is expected to have particular weight given the rapid evolution of global AI regulation. Canada's relationships with the United Kingdom, the European Union, Japan, South Korea, and Australia have all been deepening, and the strategy is expected to commit to joint work on safety standards, model evaluations, and red teaming for the most powerful systems. The G7 work plan, on which Canada continues to have significant influence, has been a useful forum for that international coordination.

The United States poses a particular challenge. The Trump administration has signalled scepticism of multilateral AI safety governance and has pursued a more bilateral, transactional approach. Canada's strategy needs to find space to maintain alignment with allies who share Canadian values while also engaging constructively with the United States on the practical questions of trans border data, compute, and talent flows that drive much of the AI economy.

What's next

The strategy itself will be released in stages. The main framework document is expected first, followed by detailed annexes on workforce, regulation, and sovereign compute. Solomon has indicated that the government will move on at least one piece of dedicated AI legislation in the fall, building on the now-defunct Artificial Intelligence and Data Act that died on the order paper before the April 2026 election.

The Bank of Canada is expected to weigh in separately, having signalled earlier this year that it intends to study the macroeconomic implications of AI adoption on inflation and productivity. Provincial governments, particularly Ontario and Quebec, are likely to release their own complementary strategies, especially given that both provinces have significant AI research clusters and are racing for federal infrastructure dollars. With the strategy now imminent, the test for Carney will be whether the document can hold together the coalition of researchers, businesses, workers, and civil society that he will need to make it stick.