As U.S. tariffs loom, Canada should use AI to fight back

With tariff threats remaining front and centre, Canadian leaders have a renewed motivation to transform our economy to be less reliant on the United States and focus more on innovation, particularly at a time when the U.S. is gutting its world-leading science agencies. 

While Canada is a global leader in artificial intelligence (AI) in general, it has specific advantages in the area of AI-assisted drug discovery. To benefit from this, Canada must harness and prioritize the development of large, mostly open, biomedical databases.

Canada is a powerhouse of talent, government commitment, and investment in AI, ranking 8^th in AI capacity worldwide. We are weak, however, in translating those attributes into domestic economic assets. It has been faster and more lucrative for Canadian companies to sell their ideas to the U.S. than build and expand in Canada. Where we need to improve is in Canada’s uncertain regulatory environment, data infrastructure, and in Canadian adoption of AI. We are making progress but need to do more.

Only 13.8 per cent of drugs  tested on humans make it to the pharmacy. Most drug candidates do not even get as far as being tested. AI-assisted drug discovery promises to dramatically change those odds.  

AI is expected to eventually reduce costs, increase patient safety, and advance drugs for the 95 per cent of rare diseases that have no existing treatment. But this is the future; today the field remains in its infancy. No country currently leads in AI-assisted drug discovery and development, offering an opportunity. 

Access to data drives AI-drug discovery and development. Data is essential for both training AI – the Nobel winning AlphaFold2 researchers drew on 50 years of well-curated and open data in the Protein Data Bank — and for testing new AI methods. Building large, well-constructed data sets takes effort, investment, and a community that works together. Testing AI predictions takes additional investment.

Canada is already a leader in leveraging biomedical data. Canadian non-profit Conscience’s Critical Assessment of Computation Hit-finding Experiments (CACHE) is a competition to test AI-predicted drugs in the Toronto labs at the Structural Genomics Consortium. 

The Artificial Intelligence-Ready CHEmiCal Knowledge base (AIRCHECK) is a Canadian-based biomedical data bank formed to support AI-assisted drug discovery. B.C. Cancer administers the B.C. Cancer Registry that it uses to partner with companies to advance drug discovery. Harnessing biomedical data — a combination of research data and patient data — would provide Canadian AI companies with an even greater advantage.

Governments, universities, and firms need to collaborate to make Canada a biomedical data leader. Similar to lifting interprovincial trade restrictions, we need to eliminate provincial regulatory differences and standardize electronic health records. 

Investments in producing and sharing large research datasets and supporting AI firms in producing synthetic data would help to achieve leadership in the field, especially at a time when the U.S.’s National Science Foundation is laying off its cutting edge AI researchers. The Canadian government has committed to investing in data infrastructure, and this is a step forward.  

Just as the AI industry has moved to open-source models of development, so too must the data be open, except to the extent of protecting privacy. Meta’s chief AI scientist said the lesson to be drawn by DeepSeek’s performance in rivaling ChatGPT was that “open source models are surpassing proprietary ones.” Groq CEO Jonathan Ross noted at the World Economic Forum that “we cannot do closed models anymore and be competitive. Open always wins,” he says.

In a world of fast-paced, open, AI development, access to data is key. Canada’s public health systems provide the opportunity for a competitive advantage over the U.S. in terms of accessing health data. Rendering Canada as a data hub will attract researchers and firms to Canada as they will want access not only to the data, but the people who constructed it and those who understand it.

Investments today in digitizing health records, reducing interprovincial barriers, creating an AI regulatory environment, and building datasets, have the potential to render Canada a leader in AI-assisted drug discovery. 

It has become clear that we must transform our innovation system to be less dependent on the US. Canada has a clear opportunity to do so with its deep scientific and research capabilities. 

Now is the time for action.

Richard Gold practiced intellectual property and commercial law at a leading Canadian law firm before becoming a professor and international advisor of life sciences innovation. He is the director of McGill University’s Centre for Intellectual Property Policy, chief policy and partnerships officer at Conscience, an independent organization building open science drug discovery to address unmet health needs, and senior fellow at the Centre for International Governance Innovation.