Canadians helped pioneer the field of artificial intelligence (AI). Researchers like Geoffrey Hinton of the University of Toronto and Yoshua Bengio of Université de Montréal—known as the ‘godfathers of AI’—laid the groundwork for technologies now reshaping economies and geopolitics. Yet as AI and the race for data become the new frontier of national power, Canada finds itself fighting to keep pace.
At the ALL IN 2025 Conference, government officials, researchers, and tech leaders gathered to confront a pressing question: Can Canada reclaim its AI edge while keeping its data—and its values—under Canadian control? The conclusion: Probably not.
It is no secret that the United States leads by wide margins in AI industry research, notable models, compute capacity, and capital. While participating in a panel at the conference, Aiden Gomez, co-founder and CEO at Cohere, commented that Canada’s contributions to the field were notable nonetheless.
“Canada led the development of this technology. We were the first ones to invest in the research when no one else believed in it. There was Geoff, there was Yoshua,” Gomez said.
In a press conference at ALL IN, Evan Solomon, Minister of Artificial Intelligence and Digital Innovation, preached about the current needs for innovation and turning ideas into businesses. Solomon specifically referred to Martha White, a University of Alberta machine-learning professor who raised 7 million CAD in her first round using AI for global water treatment solutions. Is this enough to compete with the American AI industry?
Other countries have already begun implementing AI to make government processes more efficient. In a keynote speech, His Excellency Mohamed Bin Taliah, Chief of Government Services of the United Arab Emirates (UAE)—the country of honour of the conference—explained how his government is making the most of AI.
“94 per cent of our court proceedings are all done online. People do not need to visit courts. [….] We use artificial intelligence here to summarize the proceedings between all parties and simplify the process on the judge to get the information and make a decision on the spot,” he said. “We have reduced 33 percent of the time taken to complete court proceedings by using artificial intelligence in courts.”
Bin Taliah also explained how the UAE government employs biometric scanning in their airports.
“People travelling through the UAE airports can get through the airport without even getting their passports out of their pockets.”
The Canadian government’s implementation to make AI more efficient looks a bit different. Lucy Hargreaves, co-founder and CEO of Build Canada, announced the release of its first “AI member of parliament.” It ingests bills that are introduced in the House of Commons as well as the Senate and analyzes them. Given that this AI MP just makes OpenAI GPT-5 calls, Canada has essentially announced employing ChatGPT in parliament to help parse bills. Minister Solomon referred to models from Cohere, a Canadian company, as “our strategically important LLM[s]” during the Canada AI Strong panel, emphasizing the need to use sovereign models and to support Canadian innovation. This is surprising given that the AI MP uses an American company’s model.
Canada also recently signed a memorandum of understanding with Canadian AI company Cohere to explore deploying their technologies within the government, aiming to expedite public service operations and build Canada’s commercial capabilities for domestic use and export.
Beyond internal governmental implementations, what is more economically relevant is the need to compete with both AI products and AI-powered products, while also maintaining data as a resource. At the Canada AI Strong panel, Minister Solomon emphasized the demand for tools to keep up with the shifting times, as well as the need for Canadians to trust the new tools being developed. He also referred to the idea of “Sovereign compute”, which would keep Canadian data under Canadian Law.
Computational sovereignty emerged as a recurring theme during the conference. Canadian companies and individuals often outsource cloud computing to American servers. For example, if someone wanted to train a model they designed but did not have the computational resources to do it, the task would often be outsourced to servers with graphics processing units (GPUs) in the United States. If compute resources in Canada were used instead, data would not need to flow across Canadian borders, making it less vulnerable to international security risks.
Minister Solomon was one of the most prominent voices on this topic, supporting innovation while implying the need for data protection. However, in some cases, innovation comes at the cost of weaker data security.
“Data is king. Whoever controls it, whoever uses it, whoever governs it will determine our collective prosperity and our security and sometimes our values. [….] The EU, a couple years ago went ahead and they tried to regulate, and they have found and they’re quite open about it that it’s done some good things, but it’s constrained a lot of innovation,” Solomon said.
Canadians are concerned about cybersecurity risks, and loss of privacy and intellectual property when it comes to AI; however, distrust in how the government uses Canadians’ data could be misplaced in comparison to how data is being used by companies abroad. During Minister Solomon’s press conference, he was asked what qualifies as sensitive personal data that must be protected versus non-sensitive personal data that can be exported as a resource and sold. While not explicitly drawing the line between sensitive and non-sensitive data, Solomon mentioned the need to make Canadians feel safe and protected.
“[It is important to make] sure that Canadians have control over their, you know, on privacy issues and our data,” Solomon said. “That doesn’t mean restricting data flows across borders. People need that.”
While Canadians’ concerns about the government’s use of their data are valid, this mistrust ignores how their data is already being employed by companies abroad. Most Canadian data already flows freely across borders—with Canadians’ consent. Every day, Canadians willingly supply personal and behavioural data to foreign platforms that train and deploy AI models at scales far beyond government oversight.
For example, Facebook’s privacy policy indicates the routine extraction of less frequently considered data for a variety of purposes, such as personalizing user experiences. It tracks created content, time spent on their products, friends/followers/connections, device characteristics and information, content users interact with and how, ‘device signals,’ location-related data, among other information. This data is transferred globally to their data centres and externally to Meta’s partners, third parties, and service providers. The average person does not think twice about accepting the user agreement and consenting to supply this data; with Canadians already having such low AI literacy, this is quite concerning.
After 2018 reports showed Cambridge Analytica used Facebook user information to build systems that profiled individual American voters, targeting them with political advertisements, Mark Zuckerberg testified at a U.S. congressional hearing. This led the Privacy Commissioner of Canada to investigate Facebook’s compliance with the Personal Information Protection and Electronic Documents Act (PIPEDA). The Commissioner concluded that simply clicking agree to lengthy terms and conditions does not imply that the average reasonable person would understand the nature and consequences of data collection.
This remains true today. It is unlikely that the average Canadian takes the time to understand how their data is being essentially fracked for use by foreign app providers with the click of an agree button. Moreover, it is unlikely that the government will control data that Canadians freely sign away as a condition of the foreign services they rely on. It is also unclear what security risks it imposes on the nation as a whole.
Ultimately, Canadians helped invent modern artificial intelligence, but now Canadian industries struggle to compete in the international market. At the same time, while policymakers emphasize sovereign data and safe innovation, Canadians continue to export their data daily through foreign platforms that operate beyond the jurisdiction of the government.
Canada’s next chapter in AI will hinge on more than regulation—it will depend on whether the country can cultivate literacy and the technical backbone to achieve independence. The challenge is reconciling the vision of digital sovereignty with the truth that accepted privacy policies do not respect borders.
