Science and Research Committee on Nov. 24th, 2025

[The Chair, Liberal, Salma Zahid]

Good morning, everybody. I call this meeting to order.

Welcome to meeting number 17 of the Standing Committee on Science and Research. The committee is meeting today to study artificial intelligence.

I would like to make a few comments for the benefit of 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 microphone, 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 to the chair.

For this first panel today, I would like to welcome our two witnesses. We are joined by Dr. Diane Gutiw, vice-president, corporate services, and AI Research Center lead, CGI Inc. By video conference, we also have Angela Adam, senior vice-president, sales, marketing and government relations, eStruxture Data Centers.

Both witnesses will have five minutes for their opening remarks, and then we will get into rounds of questioning.

Welcome to the committee.

We will start with Dr. Gutiw.

You will have five minutes for your opening remarks. Please go ahead.

[Diane Gutiw, Vice-President, Corporate Services, and AI Research Center Lead, CGI Inc.]

Thank you, Madam Chair and esteemed members of the Standing Committee on Science and Research, for the opportunity to provide some insights.

My name is Diane Gutiw. I am a vice-president at CGI as well as a co-chair on Canada's strategic AI council. The following remarks draw upon my 30 years of experience in applied AI research as well as implementation of data analytics and machine learning artificial intelligence in both the private and public sectors.

AI research has deep roots in Canada, driven by the early work of Geoffrey Hinton, Yoshua Bengio and Richard Sutton, which positioned Canada as a leader in deep learning in the 2010s. The country's major AI institutes, Mila, the Vector Institute, and AMII, alongside our industry-funded co-chairs in academic institutions, have placed Canada as a leader in AI research.

Your questions today were around fundamental AI research and applied AI research. Fundamental AI research has most recently focused on improving performance and making systems faster, more efficient and better at complex reasoning, and a lot of focus has gone into memory-driven architectures and multi-model systems to handle more sophisticated tasks and workflows.

Applied AI research has progressed to better analyze and generate data, text and images that enable personalized services, better decision-making, leveraging of evidence and more intelligent automation and efficiency. An example of this would be CGI's digital triplets that allow end-users, through natural language, to interrogate data from performance data down to assets and molecular-level data, which really is moving the mark on how we leverage data to make decisions.

Canada has a lot to be proud of in the areas of fundamental and applied AI research. Mila's Yoshua Bengio and Canada's CIFAR AI chair, Doina Precup, authored the first GFlowNets paper and approach for drug molecule discovery. Canada's TRIUMF and Perimeter Institute's quantum-assisted AI model combined deep learning with quantum computing to simulate particle collisions for CERN's Large Hadron Collider, and AMII's Richard Sutton won the 2024 Turing Award with Andrew Barto for conceptual and algorithmic foundations of reinforcement learning.

At the same time as these successes, the path from research to commercialization needs strengthening. By focusing investment in high-value sectors and closing the gaps in the global AI ecosystem, Canada can move to lead in translating research into real-world impact. Research institutions face slow, rigid funding, limited data access, policy barriers to collaboration and intense competition for talent. Canada must shift its culture from risk aversion to responsible innovation. Building public trust requires AI literacy, equitable access to tools, greater transparency in data use and protection, and an innovation culture grounded in security, fairness, sustainability and reliability.

Current funding models are slow, misaligned and unclear, making it difficult to scale promising ideas from research through to commercialization. Canada needs to invest in high-potential innovations beyond early research, strengthen collaboration between academia, government and industry to accelerate development, and create more opportunities for talent downstream. It's also important to ensure that our work permits, permanent residency pathways and student visa policies are aligned to attract and retain high-quality talent.

Your next question was on the role of the federal government. The federal government must play a central role in reducing risk aversion, building public trust, and enabling adoption and smooth pathways to success. To keep Canada's research competitive and ethical, clear governance is needed that aligns with Canadian culture and is supported by AI literacy. Federal oversight, strategic investments and supportive policies are vital for driving innovation and building trust in responsible AI.

The federal government should also clarify the definition of sovereignty. We need to clearly define what we must absolutely protect, such as data, intellectual property and our talent; what we're willing to give up if we collaborate and for what benefit to Canada; and who we're willing to collaborate with, and which sector...that shares our culture and ethical standards.

Looking ahead, maximizing the value of innovation requires mechanisms to bridge research to market, ensure public benefit from data assets and investments, and protect our intellectual property, providing clear opportunities downstream for our talent in academics, industry or the public sector, the areas where we need focus. Diffusion and sharing require stronger collaboration between academic industry and public sector commercialization and valorization where industry investment is key to growth, but Canada must ensure that its IP and talent stay local as well as protect our data. The practical application of innovations will help streamline the transition from research to commercialization and can help solve real-world problems and marketable problems faster.

Your final question was on the protection of public assets. Public-private partnerships are critical and we need—

[The Chair, Liberal, Salma Zahid]

Please wind up in the next 10 to 15 seconds.

[Diane Gutiw]

We need clear benefit-sharing rules to protect Canadian assets while attracting industry investment in Canada.

Thank you very much.

[The Chair, Liberal, Salma Zahid]

Now we will proceed to our second witness.

Ms. Adam, please go ahead. You will have five minutes for your opening remarks.

[Angela Adam, Senior Vice-President, Sales, Marketing and Government Relations, eStruxture Data Centers Inc.]

Good morning, Chair and honourable members of the committee. Thank you for the invitation to appear today. It's a privilege to contribute to this important study on the future of artificial intelligence in Canada.

My name is Angela Adam, and I serve as senior VP at eStruxture Data Centers. We're the largest Canadian-owned and operated data centre platform. Our facilities host some of the country's most critical digital workloads, including financial institutions, telecom providers, AI research institutes and hundreds of Canadian organizations that rely on secure, resilient and sustainable infrastructure to operate and innovate every day.

My comments today will focus on one central idea: digital infrastructure. This is my area of expertise.

Canada cannot lead in AI research, commercializations or responsible deployment unless our research institutes and universities have priority access to sovereign, high-performance and scalable digital infrastructure built and controlled here in Canada. This is a missing piece in our national AI architecture, and it is the foundation upon which every other ambition in this space must rest.

AI has advanced faster than any other previous technology cycle. Models that required modest computing power only a few years ago now demand clusters measured in tens and sometimes hundreds of megawatts. Research teams across Canada, from Mila to Vector to universities, all say the same thing: They cannot access enough high-performance compute to train and refine cutting-edge models in Canada.

We know this because some of them are our own customers, and the results are predictable. Our best researchers increasingly rely on foreign infrastructure; promising research is delayed, fragmented or simply cannot be attempted; and opportunities for commercialization drift southward or overseas. Canada has extraordinary talent, but without sovereign compute and digital infrastructure, talent cannot translate into scalable innovation.

Sovereign digital infrastructure means three things: control, where Canadian entities govern physical infrastructure, data flows, supply chains and security protocols; proximity, where researchers access high-performance compute without latency or export restrictions; and scalability, where infrastructure grows with demand, not behind it.

At eStruxture, we see first-hand how quickly this requirement is accelerating. We operate 14 facilities across the country with two others set to come online in the fall of 2026, which will be some of the largest AI-ready data centres in Canada. We're already working with leading researchers, private sector innovators and government stakeholders, who need secure and high-density environments that are sustainable and built in a way they can trust.

Building this capability is entirely feasible in Canada. We have the power, the land, the engineering talent and the climate advantages that other countries would envy. What we require is strategic alignment and predictable policy.

One of the most important considerations for the committee is that research institutes cannot compete with global cloud providers for capacity. When a global AI company wants 50 or 100 megawatts of power for its workloads, it can move markets instantly. Universities cannot and neither can our research institutes.

If we want Canadian research to stay competitive, we need to reserve a portion of sovereign AI-ready capacity for research, universities and public institutes. We need to ensure that this capacity is built on Canadian soil and governed under Canadian jurisdiction. We need to make it affordable and predictable over time, and we need to design procurement and partnership models that favour long-term national outcomes over short-term transactional decisions.

The alternative is simple. Canada will continue to produce world-class ideas that must be built, trained and commercialized elsewhere.

Data centres are the backbone of responsible, scalable AI, and Canada is well positioned to lead. A responsible AI ecosystem requires responsible physical infrastructure. As I said, in Canada, we have abundant power, and a significant amount is clean and carbon-neutral. We have strong privacy laws in an industry that is already regulated to the highest global standards of physical security and cybersecurity.

A sovereign data centre footprint would enable secure model training and storage under Canadian jurisdiction; controlled access for sensitive research; defence use cases; public sector AI; and infrastructure that can grow rapidly without compromising reliability. This is precisely the model that other nations, including the U.S., the U.K., France and Japan, are now scaling at speed. Canada must do the same.

In closing, we must forge a pathway forward through partnership, clarity and national coordination. The federal government has a critical role in setting direction, reducing fragmentation and ensuring that digital infrastructure investments align with Canada's long-term interests.

Three immediate opportunities stand out. Create a coordinated national strategy for sovereign digital infrastructure with clear targets, timelines and governance. Establish a dedicated, protected capacity for research institutes, enabling them to compete globally. Partner with Canadian-owned data centre operators, cloud providers and system integrators that can deliver this infrastructure quickly at scale and within our national regulatory and security frameworks.

We've already made this proposal to the Major Projects Office. This approach does not compete with private innovation; it accelerates it.

[The Chair, Liberal, Salma Zahid]

If you can, please wind up in the next few seconds.

[Angela Adam]

Canada built one of the strongest AI research communities in the world. We now need to give that community the infrastructure required to continue leading. If we fail to invest in sovereign, scalable AI-rated digital infrastructure, we risk losing the next decade of innovation, not because of talent or ideas but simply because of infrastructure.

Thank you again for the opportunity to appear before you today.

[The Chair, Liberal, Salma Zahid]

Thank you to both our witnesses for their opening remarks.

Now we will start our first round of questioning. We will begin with MP Baldinelli for six minutes.

Please go ahead.

[Tony Baldinelli, Conservative, Niagara Falls—Niagara-on-the-Lake, ON]

Good morning, everyone. It's a pleasure to have you here. Thank you to the witnesses for being here this morning.

Ms. Gutiw, in your comments you talked about 30 years of experience, AI research and deep roots in Canada. It's almost similar to what we've learned in previous studies. We're funding research and we have the talent in Canada, yet we have some concerns with supporting the Canadian ecosystem to grow and flourish. Work needs to be done in that area.

Ms. Adam, you talked about that as well, specifically the notion of sovereignty, and you talked about the 14 facilities you have in Canada. Congratulations on that. That's a Canadian success story as we get into this field.

In December 2024, the federal government announced it was giving Cohere $240 million as part of the Canadian sovereign AI compute strategy. They, in fact, signed an MOU with them this August. What was your reaction when you found out that one of Cohere's strategic partners benefiting from Canadian taxpayers was an American company and not Canadian?

[Angela Adam]

Obviously, that MOU could have been executed with a data centre and compute provider from Canada, but that wasn't the case. Perhaps the government has learned from that and will continue to refine their approach to look at what is available in Canada, both in digital infrastructure and from a system integration perspective. Perhaps when they make this decision in support of AI development and research, they will look at the ecosystem as a whole and pick Canadian players that can bring more value to our nation.

[Tony Baldinelli, Conservative, Niagara Falls—Niagara-on-the-Lake, ON]

We're talking about the Canadian sovereign AI compute strategy, yet the $240 million was granted to Cohere. Their first action or move was to hire an American company—I believe it's called CoreWeave—to work with it. I'm not saying that's wrong, but if we're using the Canadian sovereign strategy and funds, you would think they would look to companies like eStruxture, which has 14 facilities in Canada.

Has eStruxture received any funds from the Canadian sovereign AI compute strategy, and, if so, how much?

[Angela Adam]

No, not to date.

[Tony Baldinelli, Conservative, Niagara Falls—Niagara-on-the-Lake, ON]

Have you been consulted by the federal government in the development of the AI policy?

[Angela Adam]

We have been consulted on several occasions. We've had the opportunity to sit down with ISED, Minister Solomon and Mark Schaan. I can't really say how much of our feedback and expertise have been included in developing the AI strategy, but we have had the chance to come to the table, yes.

[Tony Baldinelli, Conservative, Niagara Falls—Niagara-on-the-Lake, ON]

Okay. Thank you.

Ms. Gutiw, we talked earlier, before the hearing started. You're a member of the task force, and thank you for your participation in that. When I look at the information posted online, I believe that 11,000 people submitted briefs to the government or took part in the consultation program.

When you look at our population of 41 million people, do you think it's just an issue of Canadians not knowing that much about AI? Why was there a slow response, especially when the government talks about creating a task force and about sprinting to create a strategy? Only 11,000 people responded. Is it a literacy issue?

[Diane Gutiw]

I believe that AI literacy is critical. It's not just a misunderstanding of what AI is. It's also a misunderstanding of what it is not. I believe this was one of the highest public consultation responses, which speaks to Canadians feeding into the request for public consultations.

I think we need to increase our AI literacy so we have better trust and comfort in what the tools are able to do and so we level set that this is a tool designed to assist us. Having more of a voice in how, in our daily life and work lives, it's able to assist us as Canadians I think is really critical.

[Tony Baldinelli, Conservative, Niagara Falls—Niagara-on-the-Lake, ON]

When we look at our previous studies, we see that we're great at funding and we're great at the talent here in the country, but it's about bringing the ecosystem to the next step of the commercialization aspect, like, for example, ChatGPT and OpenAI. It's like the private sector in the United States is moving forward in creating these systems and putting processes in place. How can we facilitate that here in Canada without always having to rely on the government to do it?

[Diane Gutiw]

That's an excellent question. Thank you very much.

I would suggest that this is exactly why we need collaboration among academic, private sector and public sector to advance it. Canada has a lot of thought leadership that has initiated a lot of the technologies and foundation for the technologies that these private sector hyperscaler organizations are advancing, and I think we're at the cusp of an excellent opportunity.

I don't think that being thoughtful in how we move forward has really set us behind. There are still a number of gaps in the AI ecosystem where Canada can thrive and become an absolute leader in everything from green AI to sustainable models to quantum computing, as well as really focusing on some niche or more—

[The Chair, Liberal, Salma Zahid]

I'm sorry for interrupting, but the time is up for MP Baldinelli.

We will now proceed to MP Jaczek for six minutes.

Please go ahead.

[Helena Jaczek, Liberal, Markham—Stouffville, ON]

Thank you, Madam Chair.

Thank you to both witnesses for coming here today.

Dr. Gutiw, I recently visited AMII in Edmonton. In the presentation that was done, we moved from researcher to researcher to learn about their individual initiatives. I was somewhat struck that there doesn't seem to be any coordination among the various researchers on the bigger questions that needed to be answered and that perhaps were priorities for Canada.

As an example, there was a demonstration of an ultrasound wand that could be used in remote communities, not by physicians, necessarily, but by nurses and allied health professionals, thus avoiding potential transfer to a southern medical centre and therefore obviously saving some money.

What coordination is going on across Canada between all these individual researchers? How do you see them best being brought together? What role could the federal government play in that?

[Diane Gutiw]

Curiosity-based research remains a critical area. I believe that organizations like AMII, Mila, the Vector Institute and others foster the ability to bring in these new innovations. We need that curiosity-based research. Where we seem to have a gap is in taking the individual research and then progressing it to the tangible, the practical. The example you have is an excellent example of how there is a need as well as a problem that can be solved with research.

As we continue to incubate these solutions that come from research, we need to focus on what problem we're trying to solve. How can we bring in the best of the industry thought leaders across Canada from the data infrastructure and from the AI compute capacity, as well as creating tangible solutions that have a benefit to society and are marketable?

By bringing together clusters of industry thought leaders to move these things along, by streamlining a lot of our great funding programs—which are hard to access, hard to understand and unclear to researchers and start-ups—and by having a single pane of glass view, how can I take this idea quickly while it is still in innovation and move it to something that's a tangible benefit?

I believe that by having groups of sectors together—health, social sciences, manufacturing, robotics, oil and gas, energy, critical minerals—we'd be able to bring in thought leaders from across a spectrum who can take this curiosity-based research, move it forward and progress it into something that will benefit us in society and be a revenue-generating, marketable solution.

[Helena Jaczek, Liberal, Markham—Stouffville, ON]

As a physician, I know that you have particular expertise in the medical information sector. Could you detail for us some of the most promising applications that you see for the use of AI to assist medical and health professionals?

[Diane Gutiw]

Health has been leveraging machine learning, AI and data for decades—my entire career—in public health solutions, not only looking at how we can use data to inform what may happen in the future in disease progression, in addiction and in the opioid crisis, but also looking at the data to see how we can get ahead of that to see how we can make a change early in a cycle of addiction or homelessness so that somebody can have a better life and we can target the problem and be able to provide a solution sooner.

Other areas where we really have moved the dial with artificial intelligence are in diagnostic imaging. Artificial intelligence is a fantastic tool to see things that are very hard to see with the human eye. We're already working with Helsinki University Hospital and others on being able to detect early brain bleeds, which are almost always fatal if they are not caught early, and on being able to reduce the screening ages for cancers because we can see minute changes earlier.

As I mentioned, there are things like digital triplets. I can have a conversation with a large amount of data that's very hard, complex and expensive to pull together. I now have the ability to bring that information together with agentic AI and multi-model AI and have a conversation with data to make better decisions and connect the dots and patterns.

[Helena Jaczek, Liberal, Markham—Stouffville, ON]

How much time do I have?