Science and Research Committee on Jan. 26th, 2026

[The Chair, Liberal, Salma Zahid]

I call this meeting to order.

Welcome to meeting number 21 of the Standing Committee on Science and Research. The committee is meeting to study artificial intelligence.

Welcome back, everyone. I hope everyone had a good break.

Before we start, I would like to make a few comments for the benefit of all witnesses and members.

Please wait until I recognize you by name before speaking. For those participating by video conference, click on the microphone icon to activate your mic, and please mute yourself when you are not speaking. For those on Zoom, at the bottom of your screen you can select the appropriate channel for interpretation: floor, English or French.

I will remind you that all comments should be addressed through the chair.

With that, I would like to welcome our witnesses for the first panel. We have Dr. Pina D'Agostino, professor, Osgoode Hall Law School, York University; Dr. Arvind Gupta, professor, department of computer science, University of Toronto; and Dr. Gail Murphy, vice-president, research and innovation, University of British Columbia. Welcome.

All the witnesses will have five minutes for their opening remarks, and then we will go into rounds of questioning by members.

We will start with Dr. D'Agostino.

You have five minutes. The floor is yours.

[Pina D'Agostino, Professor, Osgoode Hall Law School, York University, As an Individual]

Thank you, Madam Chair.

Thank you to the committee for inviting me back today.

My name is Pina D'Agostino. I am a professor of law at Osgoode Hall Law School and a tier 1 York research chair. I serve as associate vice-president, research, at York University, and as the scientific director of Connected Minds, a $318-million federally funded CFREF, Canada first research excellence fund, advancing transformative AI research and technology. I am the board chair at the Ontario Centre of Innovation, and I am the founder and director of the IP Innovation Clinic, Canada's largest pro bono intellectual property clinic.

When I appeared before this committee on December 1, I argued that Canada can be a global AI leader by supporting homegrown talent and industry, commercializing our world-class research and ensuring that AI development and adoption are ethical and inclusive and respect creators' rights.

Today, I want to underscore these points specifically, in light of AI sovereignty: why Canadian control over AI, IP and other intangible assets matters for our future, why our IP ecosystem requires long-term strategic investment and coordination, and why protecting Canadian creators remains essential to our social and economic well-being. I will focus on my areas of expertise, particularly issues of ownership, access, and control over Canadian creations and inventions, otherwise known as IP and data.

There is a great deal of discussion today about sovereign AI. Policy-makers, business leaders and commentators are calling for stronger domestic AI capacity and adoption so that we don’t fall behind in an increasingly competitive global landscape. While I agree with these points, sovereign AI must mean more than capacity and adoption. At its core, sovereign AI must address questions of ownership, access, control and governance.

By this, I don't only mean tangible assets such as data centres or cloud infrastructure. I also mean intangible assets, most importantly IP and data, which are the lifeblood of the AI sector. Ownership, access and control determine who captures value, who sets the rules of use and who ultimately benefits from technological innovation. Here, Canada faces a serious and persistent challenge not limited to the AI sector.

Crucially, much of the IP that our world-class researchers and inventors generate does not remain in Canada. Instead, it flows to large multinational firms based elsewhere. Ceding this ownership has serious consequences, including reduced investment and job creation in Canada and constrained freedom to operate for Canadian entrepreneurs wanting to innovate, scale and grow at home and abroad.

The good news is that we can fix this. As I shared before in this committee during a different study in 2023 and even earlier elsewhere, a made-in-Canada strategy is threefold. We must, one, create and protect the IP and data in Canada; two, keep control of it in Canada; and three, grow and scale it through leading global companies.

Industry, universities and other innovation hubs across Canada need this, and many are already playing a leadership role. Project Arrow is an example of what it means to build in Canada in the automotive sector. It involves Canadians coming together and contributing to a made-in-Canada solution. I am pleased to be involved in these efforts on the IP front as an advisory board member.

Canada has made some important steps through programs such as ElevateIP and via the Canadian Intellectual Property Office. Provincial governments across the country have too. IPON is a case in point. However, demand for accessible IP services continues to outpace supply.

My own work makes this clear. The IP Innovation Clinic has a wait-list of clients from across the country needing assistance. From 15 years of experience here, I can tell you that demand has not diminished despite the expansion of other IP programs. If anything, it has increased as more Canadians recognize the importance of IP. These publicly funded programs, including ISED's IP clinics program, must work together to avoid duplication, maximize efficiency and provide coordinated, national IP education and financial supports that drive IP sophistication.

Finally, as noted in my last visit and from my vantage point of more than two decades of work in this area, I will say that as AI technologies are developed and adopted in Canada, we must protect creators' rights, the lifeblood of our culture and economy. Many AI systems are trained on copyright-protected materials without creators' permission and fair compensation. This practice threatens our creators' livelihoods, risking broader social and economic harms. At Connected Minds, we take this seriously and work to ensure that AI technologies advance in socially responsible ways, not extractive ways.

I will conclude by saying that to truly bolster Canadian AI sovereignty, Canada must take greater ownership and control over the AI systems and the IP and data assets on which it increasingly relies. Our future depends on it.

Thank you again for this opportunity to appear before you. I look forward to answering your questions.

[The Chair, Liberal, Salma Zahid]

Thank you, Dr. D'Agostino.

We will now proceed to Dr. Gupta.

Dr. Gupta, you will have five minutes for your opening remarks. Please go ahead.

[Arvind Gupta, Professor, Department of Computer Science, University of Toronto, As an Individual]

Thank you, Madam Chair.

Good afternoon. Thank you to the committee for having me back. It's a pleasure to appear before you today as part of your study on artificial intelligence.

As a member of Minister Solomon's AI strategy task force, my colleagues and I engaged in extensive and far-reaching consultations with research centres, universities, public institutions, industry, and thought leaders across the country. Much of what we heard informs the federal government's strategy to shape a responsible AI research ecosystem, and I look forward to sharing these insights with you today.

As the birthplace for deep learning, Canada is uniquely positioned to lead in AI research. While it is apparent that AI has the potential to fundamentally transform our economy and society, it is less certain how this will happen. We must remain a leader in developing the underlying foundational techniques, and the mechanisms to leverage these techniques, for the benefit of Canadians, while ensuring that any and all AI systems remain ethical, fair and transparent.

Canadians must see their place in the AI revolution, garnering the skills to navigate the emerging changes in industry, civil society and government. For Canada to harness the potential of AI while managing its challenges, we must demand excellence in mission-driven research and the talent to shape and translate our research into globally competitive technologies, innovations and products.

Today, I'd like to share with the committee a number of observations from my consultations with the Canadian research community and with industry writ large.

Research and talent are the bedrock of our national AI strategy. AI continues to evolve rapidly, and Canada can only benefit if we maintain our AI thought leadership. We cannot be successful without an expansive talent pool, trained to apply the latest AI techniques, helping to implement AI to address our socio-economic opportunities and challenges. This will require taking an inclusive approach to training, from ensuring widespread understanding and appreciation of AI to ensuring that the best and brightest are attracted to Canada, ready to contribute to Canadian AI excellence.

Our AI strategy must ensure that Canadians see that AI can benefit them. The time is now to develop training and skills programs that ensure Canadians have the ability to garner the skills that allow them to adopt and thrive during future AI disruptions. Canadians must have confidence that senior business leaders have necessary AI competencies to navigate these disruptions thoughtfully. Governments must ensure appropriate social safety nets to allow Canadians to adjust to these disruptions.

All levels of government must identify strategic AI priorities and ensure that resources are directed to those priorities. Rapid changes in the AI landscape, combined with geopolitical realities, mean that rather than taking a “peanut butter” approach to policy-making, government must prioritize national strengths in AI while ensuring AI is broadly diffused so that Canadians benefit.

We will need a multidisciplinary, multisectoral approach to develop multi-use AI-based systems that can address multiple socio-economic challenges. We heard that truly inspired AI solutions can be applied across multiple disciplines and economic sectors. For example, privacy-enabling techniques developed for health care can be translated into techniques and applications for public service, education or finance.

We must ensure a national focus on building and deploying AI-enabled technologies for strategic Canadian sectors. There are sectors that are critical for the future of the country, such as national defence. There are sectors where Canada is a world leader, such as resource extraction. There are sectors where Canada is uniquely positioned to succeed, such as health care, and sectors where Canada can become a critical player in the global supply chain, such as microelectronics and robotics. For these sectors to grow and become globally competitive, we must adopt a national focus that enables the development of breakthrough technologies.

We heard that too many Canadian economic strategies mimic those of other countries, especially the United States, rather than build on this country's economic realities and competitive advantages. Getting this right and bolstering our strategic strengths can significantly boost GDP in the near term, allowing for new AI-inspired industries to emerge that build a brighter economic future that benefits all Canadians.

Issues pertaining to AI sovereignty were mentioned by nearly all stakeholders. Some noted that developing sovereign compute strategies is tricky given rapid obsolescence and massive investments by the U.S. and the EU, which make it difficult for Canada to keep pace. Canada will need to be strategic in its compute investments by, for example, investing in sovereign compute for highly strategic needs such as the pursuit of advanced AI research, while utilizing shared compute for day-to-day corporate needs.

Building on this, there is widespread recognition that we need a sovereign data strategy that ensures—

[The Chair, Liberal, Salma Zahid]

Dr. Gupta, could you please wind up?

[Arvind Gupta]

Yes.

It must ensure that domestic AI models protect Canada against future global threats.

No theme was more consistently raised than the need to build public trust in Canadian policy on the ethical use of AI. Public policy prescriptions that foster fair and transparent AI systems are a must. Success will require foundational research in responsible AI, made especially important because of the fast evolution of AI techniques. Engaging our leading AI thinkers to work with government in addressing these issues will be critical for building this trust with Canadians.

I look forward to your questions and to continuing this discussion.

[The Chair, Liberal, Salma Zahid]

Thank you, Dr. Gupta.

We will now proceed to Dr. Murphy.

Dr. Murphy, you will have five minutes for your opening remarks. Please go ahead.

[Gail Murphy, Vice-President, Research and Innovation, University of British Columbia]

Thank you for inviting me to join you today.

As you've heard, I'm Gail Murphy, vice-president of research and innovation and a professor of computer science at the University of British Columbia. I currently serve as the vice-chair of the board of the Digital Research Alliance of Canada, which is a national not-for-profit organization with a mandate to provide compute, data and software for Canadian researchers. I recently had the honour of serving as a member of the Government of Canada's AI strategy task force.

UBC is the second-largest research university in Canada, with nearly 80,000 students and more than 17,000 faculty and staff at campuses in Vancouver and Kelowna and sites throughout British Columbia. Today, over 100 UBC faculty members and hundreds of students are focused on fundamental and applied AI research in the vibrant B.C. AI ecosystem, establishing and growing companies, integrating AI into the health system and furthering critical industry sectors like the life sciences, critical minerals and forestry.

AI is a research field that's moving at warp speed with new discoveries every day. Keeping apace is not only imperative; it's essential.

Only three years ago, the power of large language models born from Canadian ingenuity became evident to the public. Already, those at the forefront of AI are thinking about how to move beyond the limitations of these models. We cannot rely on letting others invent the future of AI. We need to keep inventing it here in Canada with Canadian values, but we also need to do more to ensure that these discoveries find their way into applications that benefit Canadians. Moving AI research into applications is challenging, in part because of the many pathways by which AI can be used.

One pathway is through the formation of AI-first companies that place AI at the core of their strategy and operations. For example, Variational AI, a Vancouver-based company working with UBC researchers, is focused on providing a foundational model to speed up drug discovery. AI-first companies often require significant infrastructure to build the models on which they rely.

Another pathway is when AI is embedded into products or services. For example, UBC researchers have shown how diagnosis in rural, remote and indigenous communities such as Haida Gwaii can be improved through the right combination of training, hand-held ultrasound devices, cloud platforms and AI. Making this pathway a reality requires support to bring AI experts together with domain experts.

Yet another pathway is to develop AI solutions for critical points in the supply chains of different economic sectors. For instance, UBC professor Frank Wood has created the UBC spinoff Inverted AI, which is building predictive human behaviour models for autonomous driving. Supporting these companies to be successful can require rethinking how to integrate Canadian companies into global trade supply chains.

There are four key areas in which we can start to better support the continuum and these pathways.

The first is through research. Canada can stay at the forefront of AI discovery by enhancing and expanding the AI institute model. The institutes can be enhanced to serve as hubs that bring talent from universities, colleges, technical institutes and industry together to address mission-driven, dual-purpose challenges. The institutes should be expanded to fully deploy world-class talent that is currently underutilized, including in B.C.

The second is focus. Canada should aim to focus on and accelerate specific areas of global strength. Focus could be brought through missions or challenges posed to the AI institutes. Potential areas of focus that could broadly benefit Canadians and the Canadian economy include AI for health and AI for robotics.

The third is AI supply chains. Canada should aim to globally lead in critical parts of the global AI supply chain for particular sectors. In selected areas, Canada can establish precompetitive industrial research and development centres based on successful models in other countries for different kinds of technologies. These centres accelerate industry within the country and create new companies and a highly trained workforce.

The fourth is compute infrastructure. Canada lags its peers on AI research infrastructure in multiple rankings. To support discovery and application related to Canadians, large-scale sovereign compute with access to Canadian data is needed for both researchers and SMEs. The Digital Research Alliance of Canada is well positioned to lead on this infrastructure for Canada.

I hope the committee finds these contributions to its study helpful. I thank you for the opportunity to speak with you and look forward to addressing any questions you may have.

[The Chair, Liberal, Salma Zahid]

Thank you, Dr. Murphy.

With that, we will now start our first round of questioning, with six minutes each. We will start with MP Ho.

Please go ahead. You have six minutes.

[Vincent Ho, Conservative, Richmond Hill South, ON]

My first set of questions is for Professor D'Agostino.

Welcome back to this committee. As a fellow Osgoode Hall alum who was a fellow at the IP Innovation Clinic, I think it's very special to have you here on this committee.

Professor, I'm going to start by asking a few questions about the U.S. CLOUD Act.

Could you briefly explain the scope of the U.S. CLOUD Act and how its extraterritorial provisions operate when a U.S.-based service provider such as CoreWeave holds or controls data outside of the United States?

[Pina D'Agostino]

Thank you for having me back.

I welcome this question because it speaks to the issues I just testified on, which deal with AI sovereignty—namely ownership, access and control of data. It identifies two parts. The first is the data independence that Canada needs to maintain. The second is one that maybe doesn't come to the fore when we talk about these issues. It is technological independence.

In terms of data independence, we need to ensure that Canadians hang on to what is Canadian. Data that is created and stored on Canadian soil needs to stay in Canada, unless there is a strict exception to do otherwise.

With regard to the CLOUD Act, we're not signatories to it yet; we are in negotiations. I would encourage the government to ensure that exceptions are clearly delineated. These would apply so far, it says, for law enforcement and criminal investigations, but they're all rules-based, so the workings are as good as the system they're in. If in the obtaining of a warrant and going to court there is not a solid and robust rule-of-law-based system, that would raise concerns in obtaining data.

[Vincent Ho, Conservative, Richmond Hill South, ON]

Professor, you mentioned something about having the data stored in a physical location on Canadian soil. My understanding of the U.S. CLOUD Act is that it doesn't matter where the data is stored. Is that correct? The data could be stored in Canada, as with CoreWeave—it has a data centre here in southwestern Ontario—but it could still be subject to the CLOUD Act. Is that correct?

[Pina D'Agostino]

With the way it's configured now, it can be. That's why I'm suggesting that we need really robust workarounds to ensure we protect Canadians.

I would add that on the tech side, what is also concerning, in a sense.... I believe we're in a period of transition. We're all talking about how we need to enhance capacity, but this is really where the IP comes in.

Nvidia, I understand, is one of the main providers of GPUs. Arvind, this is your area—both of you. That can be problematic, because it could be that as the dominant player...it is also not a Canadian company. When we talk about ownership, access and control, the ideal would be that as we transit out of the period where we have been or are trying to build up Canadian capacity, we have Canadian suppliers.

[Vincent Ho, Conservative, Richmond Hill South, ON]

That's right. We learned in the previous committee meetings that the Liberal government had handed a large $240-million contract to a Canadian company called Cohere. That sounded great on paper, but then it used all of that money to buy data centre operating services from CoreWeave, which is a U.S.-based company subject to the U.S. CLOUD Act.

Although it has committed to building data centres here, it still exposes Canadians' data because Cohere is also a principal AI service provider to the Canadian government. It has access to Canadian government data and potentially even the data of Canadian citizens or residents of Canada.

Do you see that as an exposure...or a lack of government action to protect Canadians?

[Pina D'Agostino]

I see it as an opportunity for us to do better. I see it as an opportunity, as we are building up AI capacity, to.... It's all about now. We need to build in Canada and have things made in Canada, which I'm a big fan of. The ideal scenario would be that for all of that, Canadian players would be the tech providers, providing the cloud infrastructure and also the services. I would encourage us to move in that fashion.

We are in a transit period. I don't know why they were selected. It seems that perhaps there wasn't the technology or the know-how in Canada.

[Vincent Ho, Conservative, Richmond Hill South, ON]

We actually had some Canadian competitors of CoreWeave on this committee, and they said they could have done the job too had they been approached. It does scratch some heads. The Liberal government has touted this buy Canadian policy, which is great, but once again, the Liberal rhetoric doesn't match the reality Canadians are facing.

To close off, are there any other legislative gaps that you see in Canada that could help protect—

[The Chair, Liberal, Salma Zahid]

I'm sorry for interrupting, but your time is up.

I have to go to the second round.

We will now proceed to MP Deschênes-Thériault for six minutes. Please go ahead.

[Guillaume Deschênes-Thériault, Liberal, Madawaska—Restigouche, NB]

I thank the witnesses for their testimony.

Ms. Murphy and Mr. Gupta, as you know, the transformative nature of AI creates all kinds of opportunities for Canadian businesses and the economy as a whole.

Would you please talk more about how advances in AI research can be effectively translated into concrete applications and added value for Canadian businesses and various sectors?

What levers would strengthen coordination and synergy among universities, research institutes, investors, the private sector and government?

[Gail Murphy]

There are a lot of opportunities to take the research discoveries from our students or faculty at various Canadian institutions and move them to applications. Canadian industry has one of the lowest rates of adoption, as far I know, of AI in the world. We need ways to bridge the high-quality talent that's coming out of universities, to increase that talent and to provide opportunities for Canadian businesses to get their feet wet with different kinds of AI technologies.

As to the ways to do that, we can learn from other countries in how they have set up applied institutes of AI and funded different kinds of projects to give students access to companies to understand their problems. Those companies will need access to infrastructure to then try out the different AI techniques and to know they have a means of moving forward to integrate them into their businesses over the long term.

There are models we can learn from that can bring together different hubs and applied institutes, and that can let us take the fundamental discoveries that exist within Canada and move them into benefits for Canadian industry and Canadians through that.

[Arvind Gupta]

AI is still a very new field, and we're still learning what types of AI techniques we're going to need in order to be at the cutting edge. Definitely, there are many techniques being developed at the universities that can be rolled into industry.

I'll just use one of my examples here in health care. There have been huge strides made in physician-assisted diagnostics. Whether it's on the imaging or clinical side, we need to find regulatory ways to encourage these kinds of techniques to be deployed in Canada. Health care costs in Canada are skyrocketing. AI has some potential. I think there are a lot of regulatory hurdles for trying out these techniques in Canada, and we should be looking at whether we can actually take a lead there.

The flip side of this is that because AI is so new, we don't actually know all the use cases. I'm a fan of figuring out how to build a meeting ground between industry and universities so we can see new kinds of data and challenges. We have not built those kinds of meeting grounds very effectively in Canada. In Germany, they have a number of institutes—the Fraunhofers, the Max Plancks—to create a place where people can work on these problems.

It has to be both ways: Take the techniques coming out of the universities and figure out how to more rapidly deploy them in industry, and also understand problems in industry and what kinds of research can be spawned from them to develop new techniques.

[Guillaume Deschênes-Thériault, Liberal, Madawaska—Restigouche, NB]

Thank you.

Ms. D'Agostino, I looked at your online biography and noticed that you serve as an intellectual property expert with the First Nations Information Governance Centre.

Would you please explain how Canada can ensure that the perspectives and knowledge of minority groups, such as first nations or francophones in Canada, are adequately taken into account and respected in the development and deployment of AI models?

[Pina D'Agostino]

What a great question and one that's very close to my heart. In terms of the indigenous communities, the first principle that we really need to take to heart is we don't want a recolonization to happen. Whatever we do with AI, we cannot embark on an extractive approach.

I'll give you an example. Even when we're putting grants together, we often want to be engaging with indigenous communities. It's not sufficient anymore to just have a throwaway line in a grant application that says that we are going to consult with indigenous peoples. That ticking of a box just doesn't cut it anymore; we need to go way beyond that. I'm proud to say I'm living this now with Connected Minds. We have an indigenous advisory circle where anything indigenous has to go by this circle. It's an inclusive approach where they are involved in actually setting the question as opposed to projecting what we mean to study on them.

I'm really grateful for Connected Minds and the impact that we're making based on this funding, so that we now have many projects that have come to the fore, in a sense, with their own pain points. We talk about indigenous data, cultures and knowledge, which merit a special understanding and treatment that goes beyond the conventional ways that we've been used to.

[Guillaume Deschênes-Thériault, Liberal, Madawaska—Restigouche, NB]

Thank you.

[The Chair, Liberal, Salma Zahid]

MP Blanchette-Joncas, please go ahead for six minutes.

[Maxime Blanchette-Joncas, Bloc, Rimouski—La Matapédia, QC]

Thank you, Madam Chair.

I'd like to welcome the witnesses who are with us today.

Ms. Murphy, the government announced $1.7 billion in funding to attract foreign researchers. I certainly saw how the University of British Columbia responded: It seemed to welcome the news quite enthusiastically.

In contrast, however, the government is cutting funding for its own research centres. Agriculture and Agri-Food Canada is closing seven research centres, including the ones in Quebec City, Guelph and Lacombe. It is eliminating 665 positions, and over 1,000 WFA notices have been sent out.

What I'm trying to do is understand your institutional perspective. How are we supposed to attract and retain the best talent when we're dismantling public research infrastructure right here in Quebec and Canada? Some of this research infrastructure has over a century of expertise, particularly in long-term trials.