Artificial Intelligence

Honourable senators, the Senate is resolved into a Committee of the Whole in order to receive the Honourable Evan Solomon, P.C., M.P., Minister of Artificial Intelligence and Digital Innovation, to consider the subject of artificial intelligence.

Honourable senators, in a Committee of the Whole, senators shall address the chair but need not stand. Under the Rules, the speaking time is 10 minutes, including questions and answers, but, as ordered, if a senator does not use all of their time, the balance can be yielded to another senator.

I would now invite Minister Solomon to join us.

(Pursuant to the order of the Senate, the Honourable Evan Solomon was escorted to a seat in the Senate Chamber.)

Minister, welcome to the Senate.

As I have informed my colleagues, the question-and-answer period will be divided into 10-minute blocks.

These blocks will be shared between two or three senators and will include the time for your responses.

I would ask you to make your opening remarks of at most five minutes.

I would like to apologize in advance should I need to interrupt you during your interventions.

Thank you all. It’s truly a pleasure to be here. Mr. Chair, honourable senators, thank you for giving me the opportunity to speak to you today.

I appreciate the Senate’s interest in this very important work. Artificial intelligence is evolving very rapidly, and Canadians are seeing it in their daily lives — as are you — sometimes with excitement and sometimes with concern.

Before we discuss risks and opportunities, we need to talk about the practical ways that AI can change people’s lives.

As an example, last month, I was in Halifax and met Dr. Robert Chen, a pediatric cardiologist who told me up to 80% of newborns — and I have two kids of my own — are found to have heart murmurs. Families can wait up to 17 months to see a specialist, often carrying enormous anxiety even though most of these murmurs are harmless.

He thought he could help, and he thought AI could help. So, he formed an AI company and built an AI tool that could identify the most urgent cases of heart murmurs, potentially reducing the wait times, he says, from 17 months to 1 month, helping doctors focus care where it’s needed most.

This is AI at its best: a practical, humane technology that truly makes people’s lives better.

This is what AI should be about: solving real problems for real people and making sure the benefits are felt not just by a few but by everyone. “AI for all” is our guiding principle: AI to serve people and not the other way around.

Our core principle is “AI for all.”

“AI for all” means this powerful, transformative technology will work for Canadians, no matter where they live and no matter their age, income or background. It must serve people responsibly, reliably and safely, and it must be sovereign. Canadians see real promise in AI for health care, business and public services.

We have to be open to opportunities and innovation, but we must be candid about real concerns — and there are real concerns about privacy, jobs and the health of our democracy. We must and will treat these seriously. Technology moves at the speed of innovation, but adoption moves at the speed of trust, and so do our citizens.

That’s why our approach is similar.

There are three chapters.

“Protect,” “empower” and “build” are the three chapters of our upcoming national AI strategy. First is “protect.” AI will only succeed in Canada if people trust it’s being developed and deployed responsibly. That means stronger privacy protections, actions on harmful “deepfakes” and online manipulation, protecting children and vulnerable communities and building real safety capacity here at home.

That’s why we created the Canadian Artificial Intelligence Safety Institute: to get a better grasp of the risks and to put real guardrails in place.

Second is “empower.” Canadians don’t just want to be protected from AI; they want to benefit from it. They need and want the skills to use it. They want their children to be ready so there are jobs of the future. They want workers and businesses to have the tools to adapt and succeed.

That means investing in skills, digital literacy and opportunities for Canadian workers, youth and businesses.

It also means making sure AI reflects Canadian values, including with respect to inclusion, culture and Indigenous leadership.

We also have to respect our two national languages.

If people feel left behind, adoption will stall. If they feel empowered, it will grow — and our strategy will have a pillar on empowerment.

Third is “build.” Canada has extraordinary strengths in artificial intelligence. We have world-class researchers, innovative companies, clean energy, strong talent and the capacity to lead. Leadership, though — and we are leaders — is not a birthright. We need to fight for it, and that’s why we need to build. That means investing in Canadian infrastructure and sovereign compute. It means helping Canadian companies scale here at home and championing our champions. It means keeping more talent, more of our intellectual property and more value here in Canada.

Canadian AI must not be something invented here and developed elsewhere. It has to help build a stronger Canada right here at home.

Too often, we plant the seed and water and grow the plant, and then someone else harvests it. That will stop. Our plan will take that under consideration.

Thank you, minister. We will now go to the questions and answers. The first question will be from Senator Carignan.

Minister, thank you for accepting the Senate’s invitation.

My question is about Anthropic’s Claude Mythos. It’s a next-generation AI tool that’s being used to identify vulnerabilities in cybersecurity systems. I know that you received critical information over the past few days about the capabilities of Claude Mythos.

So they’re keeping Mythos under wraps. Anthropic decided not to release it publicly and to create a group called Glasswing so that cybersecurity companies can use Mythos and identify their vulnerabilities themselves. I’d like you to comment on that, minister. Have you talked to Anthropic about including Canada and its agencies in Glasswing so we can test our own cybersecurity systems?

Thank you for the question, senator.

It is a very good question, and I’m glad that you’ve asked it.

My team met with Anthropic last night. I met with senior leadership of Anthropic this morning. We had very productive discussions about a number of issues, senator, including Mythos and including Glasswing, but also about other issues that have been raised that I have promised the public I would speak about, as I spoke about with OpenAI — questions about safety, about protecting our children and consumers, about how they escalate and what their safety protocols are.

Senator, our number one goal is to make sure our citizens, our children and our institutions are safe. There are new developments almost every day. We are very much at the forefront, as you can see from my meeting with Anthropic. The conversations were very productive.

I cannot give you the details yet. I know you will ask for the details about whether we’re in. I can’t give you the details except to say that we are fully engaged in very productive conversations, as is the Canadian AI Safety Institute that we have.

We have the full measure of our systems that protect Canadian institutions that are engaged not only on this particular issue but on cybersecurity in general.

Well, there have been meetings. The chair of the U.S. Federal Reserve met with the big American banks to discuss concerns about Mythos. It is a national security issue. Do you plan to ensure that Canadian agencies and Canada will have access to Mythos as soon as possible so they can detect vulnerabilities in our security systems?

Senator, you are asking the exact right question. We share a concern to make sure that Canada is at the very forefront of our safety and our institutions.

We have followed the situation with Mythos very closely, by the way, and with many other factors. Again, I point to the fact that I had a very productive meeting with the Anthropic leadership this morning.

My team is fully engaged. Our government, as a whole, is engaged in making sure that Canadian institutions are protected from any cybersecurity threat posed by AI, including the Canadian Cybersecurity Network, as you know, senator.

I will promise you, senator, that these are ongoing discussions. We will be very much at the forefront of making sure, through these productive discussions, that Canada, citizens and our institutions — banks and our very important institutions — will have the capacity to be protected from this kind of cyberthreat.

Have you also included your colleagues from National Defence, Public Safety and Finance, given that these security issues certainly affect those departments as well, not just your department and your portfolio?

Senator, you’re asking exactly the right questions.

I can tell you that the Minister of National Defence and the agencies are engaged very closely in this. The Canadian Centre for Cyber Security and Minister McGuinty — we are working closely to make sure that we’re tied in to understand the security risks and to mitigate them.

Senator, the Canadian Centre for Cyber Security — over this weekend, as we came to ground on this very pressing issue — published guidance for businesses on this that you can look at.

I know that the Bank of Canada met with major lenders on the cyber risks related to Mythos a number of days ago.

As you know, we have engaged the Canadian AI Safety Institute to look at safety protocols with OpenAI. That institute is also a mechanism to make sure on that side as well.

I assure you and I’m happy to tell you and the honourable members here in the chamber that we are engaged in this subject. I’m working with my colleagues closely and very urgently.

We’re counting on you. You said that we are leaders in AI, that we have major research chairs, that we have some of the most important chairs right here in Canada. However, Canada is lagging behind when it comes to integrating AI into businesses and how businesses use it. This leads to a loss of productivity, first and foremost, and contributes to the brain drain, affecting both researchers and professionals in the business sector.

I was looking at the OECD figures to compare Canada with our neighbour to the south. The figures are in euros, but when it comes to research and development, the United States allocates 90 billion euros a year, while we allocate 2.5 billion euros. That is minimal compared to our investment, despite all our good intentions. What’s the plan to ensure that AI is better integrated into our businesses, first of all, and secondly, to make sure we retain our top talent, who are being offered — I say this jokingly — hockey player salaries to go to the United States?

Again, senator, first, thank you for the question.

Retaining talent is absolutely key. We are world leaders, I should say, in many aspects of AI, including in research. Just so we make sure that people appreciate, our three national AI institutes — Vector in Toronto, Mila in Montreal and Amii in Edmonton — are run by the three great godfathers of AI: Geoffrey Hinton, Yoshua Bengio and Rich Sutton. They, in turn, have attracted an enormous amount of ecosystem, and so have our CIFAR chairs, bringing world-leading PhD students, postdocs and innovators to Canada.

In terms of talent attraction, though, in our last budget, senator, we put $1.7 billion in the budget over 13 years for the international talent attraction strategy to attract 1,000 of the world’s top scientists, and their labs and students, to Canada. This is the largest talent attraction strategy in the G7. Four hundred million dollars of that goes to the Canada Foundation for Innovation to ensure that the chairs there have the equipment they need to conduct the research.

We’re making sure in our national strategy that the students, once they exit, have a path to commercialization. This is why you’ve seen in our productivity super-deduction we have the lowest marginal effective tax rate right now, lower than in the United States, attracting capital. In our national strategy, senator, we will make sure that we have attraction for making sure that risk capital is available for our start-ups.

To keep our companies here, they need the fuel for innovation, and that’s compute. That’s why we launched the $300-million AI Compute Access Fund so our small- and medium-sized enterprises have the compute capacity to stay here and grow here with Canadian companies.

Thank you, minister.

Thank you for being with us, minister.

Minister, last month, before representatives of OpenAI, you said, “. . . protect our children or the hammer comes down.”

As you know, unlike the EU AI Act, which imposes specific obligations to protect children, AI platforms in Canada are not currently subject to comparable binding legal obligations to protect children.

My question is this: Within your responsibility, what concrete measures are we taking right now to reduce harm to children and youth, while, of course, still supporting Canada’s AI development?

Thank you, Senator Petitclerc. That’s a great question.

We have three aspects to approaching harm in a whole-of-government approach. Justice Minister Sean Fraser has already put forward a tabled bill, which I hope gets passed, that will criminalize the non-consensual sharing of sexualized or “deepfake” imagery. We believe that is a form of violence — violence against women, violence against children — and we would like that bill passed in order to criminalize that. That is currently tabled legislation. That’s the justice side.

On the Minister of AI side — my department — I’m in charge of updating the Personal Information Protection and Electronic Documents Act, or PIPEDA, our 26-year-old privacy legislation. I will be putting forward legislation that will update our privacy legislation. I can’t give you every detail that’s in that, senator, but I will say three things very quickly. One, it will include protecting children and how their information is treated. We are looking very closely at things like transparency for automated decision systems. In other words, those AI and decision systems that may be used to provide a mortgage for someone or a job interview. Also, I’m looking very closely at things, senator, like the right to deletion so that people have the right to take things down. We’re looking closely at that.

There will be my privacy legislation that will come down, and then there is the online harms element that will fall more or less under the Minister of Heritage and Culture Marc Miller. That will be looking at things like online bans. But with regard to that whole-of-government approach to protecting our children from online harms, if I may, I would just say to the senator’s point about my remarks after OpenAI that, currently, the Canadian AI Safety Institute, as I’ve instructed them, is engaged with OpenAI to look closely at their systems and will report back to me so we can decide what type of regulation would be best suited.

Minister, the latest data from Statistics Canada suggests that, while aggregate job losses from AI have not yet emerged, a more concerning pattern is taking shape. Employment growth among Canadians aged 15 to 29 is lagging significantly behind older cohorts.

As part of the government’s 2025 Building Canada Strong consultation, young Canadians consistently identified the lack of entry-level jobs as their primary barrier to employment. Given that generative AI is now automating the junior task-based work that once served as the training ground for early careers, what specific steps — I underline, specific steps — are you taking in your AI strategy to preserve pathways into the workforce and prevent today’s 13.8% youth unemployment rate from becoming a permanent feature of the economy?

Thank you, senator. That is an excellent question. I am a father of two kids in their early 20s, and first‑time job applications are literally what I talk about every day with them and their friend groups. So this is a poignant conversation. Let me say a couple of things.

Let’s just get the context. We heard earlier from a senator about our adoption rates. Although Canada has great research, we lag in terms of adoption here. In the Organisation for Economic Co-operation and Development, or OECD, Canada is one of the laggards. We are adopting AI — in other words, using AI — at only 12%, so it hasn’t deeply penetrated the job force so that we can start genuinely seeing those macro impacts on jobs yet. Although, we believe, like you, that this is something we are deeply concerned about.

Let me also just outline very quickly that we have about 150,000 Canadians working right now in the pure AI sector alone. The digital innovation sector comprises 800,000 Canadians. It is the fastest-growing part of our economy. AI alone has 140,000 jobs. I will also say that Deloitte Canada and the Vector Institute have done a study on the impact of AI so far on our economy. They say that AI-related jobs have contributed $82 billion to $100 billion across Canada over the last five years alone, and AI is projected to add $298 billion to our GDP and create 41,500 new jobs annually over the next decade.

On one side, you have predictions that this will create jobs, as most technologies have. That doesn’t mean we should not make sure that we have training, and I will tell you that we are taking action with AI4ALL. As I say, one of our core pillars will be empowerment, and empowerment will be upskilling, job scaling and working with organizations across the country that are doing superb work.

For example, organizations that we’re looking at working with are those like Mitacs that do apprenticeship programs and training for young people. So we are making sure that young people have the tools, the literacy and the skill set to enter the job market and create the job market. For those who want to start companies, we need to ensure that they have access to risk capital, and for those who need the skill set to use it in manufacturing and whatever part of the job. So that is a foundational question.

Thank you, minister. Upon reflecting on the tragedy in Tumbler Ridge, you stated that artificial intelligence is not theoretical, that it has real-world consequences and that this requires responsible governance. That principle is not in dispute. We are now seeing approaches that would permit organizations to define their own standards for collection, analysis and use of personal data in AI-driven decision making, even where those decisions affect individuals’ economic opportunities and participation in society. That is not consistent with how privacy law operates elsewhere in Canada, where baseline standards are established in law and are enforceable.

Can you confirm whether the government’s position is that AI governance in Canada will be grounded in clear, uniform and enforceable legal standards, or whether it will permit continued reliance on self-defined standards in certain sectors?

Thank you for the question, senator. I want to say that I know I speak on everyone’s behalf when we think about the families in Tumbler Ridge, the tragedy that happened there and the victims. We think about those families who are going through this every day.

As you know, in that specific case, as you referenced, within 48 hours, we summoned OpenAI to Ottawa. I spoke with their senior team. I was disappointed with their self-regulation that you talked about when they said, “We have our safety protocols.” I then demanded a meeting with their CEO, Sam Altman, and got that as well. I made five important requests, including that the Canadian AI Safety Institute get a very close look at their safety protocols, because we have to understand what these technologies are doing. They are brand new.

Senator, these companies are going through 2.5 billion messages a day, and they’re doing threat assessment through technology and human intelligence. So we demanded a number of things: one, that they make sure that they have Canadian mental health experts and Canadian legal experts if they’re going to escalate; two, that they have direct contact with the RCMP. Their self-regulated system reported to the FBI, so they had to report directly to local law enforcement. They had to relook at their past year and reassess their flags. Finally, they have to open up the door to the Canadian AI Safety Institute to have a good assessment.

I will confirm to you that, no, we will not rely simply on self‑regulation. We are looking closely at the standards that we will apply. But, as you know, we have to get good information to make good law, and I will be responding when I get the assessment of the Canadian AI Safety Institute. However, you’re right, when we pass privacy legislation, we will be grounded in law.

Minister, thank you for being here. There is an issue of trust in adopting AI: literacy generally and AI literacy specifically. So far, Canadians have been slow AI adopters, and we heard during our AI study that research dollars are meagre for both trust and AI literacy.

This recent headline in the Toronto Star captured my attention:

Our leap into an AI future will be a stumble, if we don’t fix this one big problem in Canada.

It refers to literacy generally and AI literacy specifically.

Minister, how do you intend to close the widening literacy gap — as the government and employers rapidly embed AI into everyday decisions — without leaving many Canadians behind?

Thank you for your excellent question, Senator Burey.

It goes, again, to what I was answering in response to Senator Mohamed as well. Skills training and literacy are absolutely key. We fund Amii, one of the national AI institutes, as an example. They currently offer a literacy course for students, and they are closing in on — don’t quote me on the exact number — almost 800,000 Canadians whom they hope to reach this year on AI literacy. These are three-hour courses just to give basic AI literacy. There are others as well.

You started with safety and trust. There’s skills training, but there’s also trust. I should say that there’s literacy for students and also for businesses. We have our voluntary code of conduct that 48 different organizations have signed onto, including big organizations such as IBM, TELUS and Cohere. Although it is a non-binding agreement regarding guardrails on safety, transparency, fairness and equity and oversight, we are outlining our policy that you will see in our national AI strategy, which will comply with our new privacy legislation. But, again, in our national AI strategy, you will see that AI literacy and skills training will be a core pillar.

Thank you.

Minister, thank you for appearing before us today and for your opening statement. I personally commend you for your dedication, devotion, in-depth engagement and command of this file.

Minister, could you tell us why you are drafting a new national AI strategy and why it is so important for Canada at this point? Also, given that you have talked numerous times about the importance of supporting Canadian companies in this domain, what has your government done to bolster our innovation and AI champions?

Thank you. This is a critical moment. Senator, your question is exactly right about the need to renew our national AI strategy. Canada was the first country to have an AI strategy back in 2017, but we will be renewing it almost a year and a half sooner than originally predicted because things have changed. The world has changed, as the Prime Minister has outlined.

We have a geopolitical realignment at the same time as a technological realignment. Given the Prime Minister has called this a “hinge moment” — I’ve called it a “Gutenberg moment” — we need to renew our efforts to redouble our investment to champion Canadian champions, to educate our workforce with skills and to ensure we protect our consumers, our privacy, our data, our jobs and our government-sensitive data as well. Citizens are concerned about that. They are concerned about ensuring that we do not rely on the technology of other countries. As the Prime Minister said, we do not want to choose between hegemons and hyperscalers. We need technology that is safe and sovereign under Canadian law, free from the coercion of other countries. That is the kind of sovereignty of sensitive data that Canadians demand, and we’ll be building that.

This is why we formed our 28-member national AI Strategy Task Force. This is why we had 11,000 Canadians — more than any other engagement in Innovation, Science and Economic Development Canada’s history of public engagement. We’ve published those for the public to see. There are over 11,000 comments — not little comments, but long comments. It’s a half‑hour form. So it’s important that we renew the strategy.

In terms of championing our champions, we have national champions. You have now seen our government ensuring that we support our companies with what we see as the three Cs and the two Ts. Regarding the three Cs, this is what we heard from our task force: What do they need to thrive? They need compute, and that’s why we have the AI Compute Access Fund. They need customers. The government has to play a role on the procurement side, so that includes the “Buy Canadian” policy and the procurement, which is why we have the memorandum of understanding with Cohere and Coveo. They need access to capital. As we’ve already done in the last budget and further, you’ll see us making sure they have access to risk capital so that they can invest.

The two Ts are trust and talent, which we’ve spoken about here. Providing those will absolutely help create the sovereign AI that Canadians need and trust.

Minister, thank you for being here.

Indigenous Elders remind us that culture is in the language. I have been exploring a project with a group of national and transnational tech firms to potentially use AI for interpretation and translation between English and Mi’kmaq. I know that AI for language preservation is a topic that the National Chief has also raised with you.

Minister, I have two questions: How is your government approaching the issue of digital sovereignty, ensuring that Indigenous Peoples continue to maintain the rights to traditional and sacred knowledge that they feed into training modules such as language?

My second question is: My project would benefit greatly from recordings and archive material held across the country and around the world. How is your government working on digital repatriation to put this type of documentation and knowledge back into the hands of the people whom it was taken from before we add things like consent forms, IP protections and contracts?

Thank you.

Senator, thank you for the excellent, vital and foundational question.

When I co-hosted the G7 with Minister Joly, we had Elder Kevin Deer of Kahnawake, and he said to me, “Minister, you’re into AI. So am I.” I said, “Great.” He said, “But you’re into artificial intelligence, and I’m into ancestral intelligence. You have to make sure that your artificial intelligence guards, upholds and protects our ancestral intelligence.”

It was such a powerful moment. I’ve spoken to Kevin many times since. In fact, I offered him tobacco so that I could borrow that phrase, and he accepted.

It’s vital. The ancestral intelligence of Indigenous Peoples will be a vital part of what we do. By the way, I made sure Elder Deer spoke to the entire G7 to say that.

Now let me get to what we will do. Language models are absolutely key. In fact, I spoke to Premier Wab Kinew again today. He was in town.

Premier Kinew is very seized with this. He has a program where he’s building a repository of Indigenous languages based on the Hansard from the Manitoba legislature, translating it — because they are huge databases — in order to make sure that AI can train on Indigenous languages. We need data sets. We have offered to support that.

I have met with the National Chief, whom I connected with again last night, and many Indigenous leaders about this very issue. We are seized with it. It will be a key component in our national strategy.

I will also say this: We have met with many Indigenous leaders about data centres. This is very important. Many nations are very involved in terms of the location of data centres, water usage and land rights. These issues are foundational to what we are doing, and our consultation continues to be very extensive on that.

Thank you, senator.

I worry about sovereignty and governance a lot, and I know that we cannot sit on the sidelines, that a steady state will, in fact, not only abdicate our leadership but also put our sovereignty at risk.

As well, end-to-end sovereignty may not be viable for many countries, Canada possibly included. We know that data travels and often boomerangs outside Canada before it lands in Canada for storage.

We know that data is not necessarily under Canadian control when it sits on a foreign-owned cloud by a foreign-owned company, even on our soil. You’ve spoken about sovereignty, not solitude.

I believe you’ve signalled openness to U.S. companies on a hybrid, sovereign cloud; we likely have to do that.

What specific legal mechanisms, governance and infrastructure will ensure Canadian data is protected from foreign reach, such as instruments like the U.S. CLOUD Act?

Senator Hay, first of all, thank you for the work you’ve done on AI education. It was a great pleasure and an honour to appear with you about that.

You’re asking an incredibly important question. We hear about the CLOUD Act, the PATRIOT Act and about Canadians wanting to make sure their information is protected.

Let me just be as clear as I can with everybody here. Sensitive, important government data will be controlled under Canadian law in Canadian data centres, full stop. This is really important for sovereign data.

Data sovereignty is also not just about storing data here in Canada. It is not enough to have the data centre located here, because if it’s stored in an American cloud, it could be subject to the CLOUD Act.

Let’s be clear: The CLOUD Act is also subject to warrants. It is very specific but, nonetheless, subject to foreign law.

So I will say that we will make sure that in our strategy — and we are currently doing this — we build up not only our sovereign Canadian companies but also sovereign cloud.

Now, you mentioned hybrid. There are many levels of data that people will protect at different levels. Very sensitive data will require a fully Canadian sovereign cloud, and that will be for government choices.

For other levels of data, there may be different models, what’s called “two-key security,” where the data is inaccessible even if the CLOUD Act were invoked. If the two-key security is encrypted, the data would still not be available, but, at the same time, you might lose access to the tools to use that data.

The issue of sovereignty is very important for sensitive workloads, and we’ll be working with that. Again, for the most sensitive government service data, I just want to say that our standard is clear: Canadian law, Canadian governance, hosted on Canadian-resident infrastructure, secured with strong technical protections.

Minister Solomon, tomorrow I will be speaking to 500 participants at the National Center for Trauma Informed Practices Conference with its executive director, Kevin Cameron.

Mr. Cameron, an Order of Canada recipient, was responsible for developing the Traumatic Event Systems Model following the Taber school shooting in 1999. There are now thousands of educators, mental health professionals and police officers trained to assess potential threats to use violence among youth in Canada.

Risk assessment of online behaviour has proven to be complex and difficult to detect before problematic behaviour occurs, and we already know the impact that AI chatbots are having on vulnerable youth.

Minister, what can I tell these 500 participants to give them hope that the government is addressing the online safety of youth and preventing the manipulation of youth through AI chatbots, including tools for professionals to assess such risk?

First of all, thank you for the question. Again, safety underpins all the work we have done and will continue to do in the AI strategy. This is why we have the Canadian Artificial Intelligence Safety Institute, and we are taking these approaches very seriously.

First and foremost, we’ve already tabled, through the Minister of Justice, as I’ve said, legislation to criminalize the non‑consensual sharing of deepfakes. This is really important. We hope that passes because that is a form of violence against very vulnerable communities.

In our upcoming renewal of the privacy legislation, as I said publicly when we tabled it, I’m interested in taking stronger measures. We’re looking at things like the right to deletion that would allow people to take down deepfakes, or synthetic imagery that is harmful, with penalties.

In terms of online harms, again, I don’t want to stray too far into the work of my colleague Marc Miller, but I’m very engaged with him on that. We are looking at practices that have occurred in places like Australia with social media bans under a certain age.

There are questions, senator, we have to come to ground on, which we are looking at closely, regarding age verification. How do we verify the age of children without giving the very information we’re protecting to the people we’re trying to protect them from? Age verification, which can be either biometrics or key data, is very important.

Now, there are some solutions we are looking at. They exist. We know about them. We’re looking at them. But verification is a little more complicated than enforcing it. And what’s the enforcement?

I don’t know if I have time — I don’t. It’s a great question.

Minister, in Budget 2022, Canada launched a National Quantum Strategy, building on major investments it made in quantum research and technology development hubs through the Canada First Research Excellence Fund.

From this funding, UBC, Sherbrooke and Waterloo have built impressive co-labs and are now joined by Calgary in that endeavour. However, funding is now running out. The federal government has prioritized these technologies for economic growth and national sovereignty.

How does the government plan to ensure Canada maintains global research and talent development leadership in quantum technologies, which is required to continue to grow the quantum economy in Canada?

My second question: Vancouver, where I’m from, is home to a world-class AI research and innovation ecosystem. By any measure, Vancouver is one of Canada’s global centres for excellence in AI research.

However, since its launch in 2017, the government’s Pan‑Canadian Artificial Intelligence Strategy has not adequately recognized or leveraged the strength of B.C.’s leading research universities, vibrant ecosystems of small- and medium-sized enterprises, SMEs, and major global companies. Will the minister’s forthcoming AI strategy address that gap?

Thank you, senator. Great questions.

Let me talk about quantum briefly, and then I’ll talk about the marvellous ecosystem in British Columbia.

First of all, quantum is absolutely important. We have over 100 companies here in Canada. The United States Defense Advanced Research Projects Agency, DARPA, program had a Quantum Benchmarking Initiative program to try to see if they could create a fault-tolerant quantum computer, which is very important because, if you can create one, you can break through forms of security. From a cryptography point of view, all your security can be broken with this type of computer, so it’s very important.

They created a very lucrative challenge that some core Canadian companies were involved in. That’s why in December we launched the Canadian Quantum Champions Program. Stage 1 of our program is more lucrative than their stages 1 and 2 combined.

They have a $300 million offer to any company that hits their stage 3 and then another US$300 million to move their headquarters to the United States. We have kept them here. Our four core companies are involved, and more will be involved in stage 1.

Our goal is to make sure that we keep our quantum companies here for security, for intellectual property, IP. And our Canadian Quantum Champions Program is doing just that, rivalling the U.S. program. In fact, it’s more lucrative in the early stages and will rival it as we roll out.

Any time a company accepts money, the goal is to make sure that their headquarters, IP and ownership stay in Canada to keep our IP here.

We also have the National Quantum Strategy to commercialize and develop quantum, with about $360 million attached. We have done a lot on quantum in a very short period, and there is more to come in the national strategy.

With my remaining time, I will say we are deeply engaged with the ecosystem in B.C. We are deeply engaged with the government there to make sure that we capitalize on this remarkable ecosystem in physical AI, robotics, health care, life sciences and technology to ensure that we benefit from that —

Thank you, minister. We will now proceed with the next block.

Minister, when you were appointed as Minister of Artificial Intelligence and Digital Innovation a year ago, Prime Minister Carney did not give you a mandate letter specific to your brand new ministry. Instead, the PM only gave you and all of your cabinet colleagues the same general, vague two-and-a-half-page mandate letter. That was not really a businesslike approach from this government.

Without a specified mandate for your newly created cabinet portfolio, how can the Canadian public know what the Prime Minister expects you to achieve in this position? Your general mandate letter does say this:

Over the coming weeks, I will look to each of you to identify the key goals and measures of success on which to evaluate the results you will achieve for Canadians as a member of the Ministry.

Since you received that letter almost a year ago, what was your response to the PM? Please provide us with a copy of your response, as it should be available to Parliament and the Canadian public.

Senator, thank you for the question. We’re very proud that, first of all, the Prime Minister created the first Ministry of Artificial Intelligence and Digital Innovation, recognizing the importance of focusing on this issue and the impact it’s going to have on our economy and our country.

The mandate letters were a reflection of the Prime Minister’s approach, which he was very transparent about, and the key goals that we have as a government to build, empower and protect. We, as ministers, have responded to that.

You will see in our national strategy that part of my mandate is to renew that national strategy early. Senator, you have seen a number of strategies — including the Defence Industrial Strategy, an auto strategy and a nature strategy. Very soon, you will see a national AI strategy that will respond to core concerns — and not just concerns but, as I outlined in my opening remarks, the mandate to ensure we’re building in Canada with sovereign AI, protecting Canadian citizens but making sure we take a leadership stake in the world.

You will see that with our national strategy and, of course, in renewing our legislation to finally, we hope, pass a renewal and update of our privacy laws.

I would like to receive a copy of that response to the Prime Minister. I’m sure all parliamentarians would like to receive that as well. I’ll move on, though.

There are only two vague mentions of artificial intelligence in that entire general mandate letter. One of those mentions says, “Government itself must become much more productive by deploying AI at scale . . .” What does the phrase “deploying AI at scale” mean? What are your government’s projections for how many federal public servants will lose their jobs after they’re replaced with AI?

I appreciate the question. First, it is absolutely imperative as a government — and I know you share this concern, senator — to serve Canadians better. Our job is to serve Canadians better and to ensure we’re doing so in as efficient a way as possible so money goes to Canadians, not to the operations of government. That is exactly why the Prime Minister not only appointed me but also appointed Joël Lightbound as the Minister of Government Transformation. It is exactly why he has talked about setting up an office of government transformation. It is exactly why we have made sure we have procurement policies to ensure we buy Canadian and buy Canadian technology. It is also exactly why I did an MOU, signed not only by me but by Joël Lightbound and the Treasury Board, to make sure that we are using strategic companies like Cohere or Coveo — and there will be others. This is to ensure we’re using the best tools we have, according to Treasury Board regulations, in a safe, secure and reliable manner, while ensuring we’re serving Canadians as efficiently as possible using Canadian technology and supporting Canadian innovation.

That’s why you’ve seen things like CANChat, GCtranslate and other aspects where we are making sure that we’re using the technology in a way that enhances service for all Canadians.

If you don’t have those projections to provide today, I’m sure your government does. I think federal public servants and Canadians deserve an answer about how many of them could potentially lose their jobs because of AI.

Also, I could not find a lot of information about what underpins your brand new ministry. It seems as if you do not have a deputy minister per se but instead an official who is associated with the Privy Council Office. You apparently receive some support from the Department of Industry. However, it appears that there are only a few bureaucrats directly associated with you. Meanwhile, it also appears you have about 17 staff in your ministerial office.

Minister, is it true that your ministerial office staff potentially outnumber by several times a very small bureaucracy that’s actually assigned directly to you?

Thank you. Senator, I am delighted to hear that you are expressing concern that there are too few bureaucrats working and that we’re perhaps not using resources and technology to be more efficient.

We are working together closely. This ministry falls under Innovation, Science and Economic Development Canada, or ISED, as well as the Minister of Industry and the Minister of AI and Digital Innovation. We are served by a remarkable team of public servants who have given us full support. Here with me today is Mark Schaan, deputy minister. I assure you we have access to and full support from a remarkable group of hard‑working, intelligent, innovative public servants who have given us nothing short of superb service. Also, the Clerk of the Privy Council, Michael Sabia, has deep experience in not only making sure that we are getting the best out of our public servants but also that the best public servants are doing the job of serving Canadians.

I also have confidence in my political staff, whom you mentioned and who are doing remarkable work. Again, I don’t in any way want to disparage the work of public servants, who do a phenomenal job.

Also, you can do innovative, superb work and use innovative technology. Canadians expect the government to walk the walk — to use the tools that we’re talking about to make sure they’re efficient. Our public servants are open, willing and innovating using those same tools to serve Canadians better.

Thank you, minister. The last three minutes of this block will go to Senator McCallum.

Thank you for being here. I have two questions.

First, there are still First Nations communities without access to reliable, secure internet services. How will the government ensure that First Nations, Métis and Inuit are not once again left behind? How will you support connectivity?

Keeping that in mind, how can Canada’s AI strategy help promote data sovereignty with principles of OCAP, or ownership, control, access and possession, regarding health care data for First Nations, Métis and Inuit peoples, regardless of where they live?

These are profoundly important questions, senator. Thank you.

First, we’ve done extensive consultation with First Nations, Inuit and Métis leaders and communities on these very questions.

You have asked a very profound question because before we can talk about AI, we have to talk about universal access and our promise as a government to provide universal broadband access to ensure that communities — wherever they are, as you say — have access. Our government has made a pledge to have 100% universal broadband access by 2028. We are seized with that because you need to have a first step.

In the meantime, we are also taking steps, as you know, to make sure we’re advancing Indigenous-led projects, Indigenous-led AI, Indigenous-led data centres and Indigenous-led models that train on languages. You’ll see in our upcoming AI strategy that it is a core point that we are making sure that we are advancing Indigenous-led technology projects.

You made a second point about safeguarding data. It’s really important. There’s a trust issue with data, and this is why sovereign Indigenous AI and sovereign AI in Canada are very important elements. It’s because there’s a trust factor regarding who holds the information. For standards and practices, as we discussed with leaders and Indigenous communities, it’s important that we work to build those capacities. It is important and essential that we engage in the hard work that that will require. Thanks.

Minister Solomon, witnesses across multiple Senate committees consistently emphasized three converging risks: one, the erosion of human rights through algorithmic bias; two, the rapid deployment of AI systems in critical infrastructure and communications without clear safety standards; and three, the growing gap between innovation and public trust — all of which we’ve discussed. In that context, is the government actively considering treating certain AI systems and capabilities as public infrastructure, akin to transportation or telecommunications, which would reflect Canadian values and be subject to public oversight, shared standards and equitable access? If so, what could that approach look like in Canada?

Thank you for the question. It’s a profound one. There’s a lot in there, senator. We’ve covered the online harms element. I don’t want to repeat the fact that we will be taking strong measures on online harms and a comprehensive approach on the online harms side, the privacy legislation side and criminalizing “deepfakes” to build trust.

Our democratic institutions have to be protected. That’s why we have Bill C-25, the strong and free elections act, to help maintain the Canada Elections Act, and that will, of course, amend, we hope, the Canada Elections Act to prohibit “deepfakes” from imitating electoral actors to mislead Canadians and limit foreign interference. So there’s a protection side that will be important.

On the public good side, first of all, currently, our $300‑million AI Compute Access Fund subsidizes access to compute for SMEs, so we’re trying to encourage that. Secondly, we are building an AI supercomputer that is funded so that our public, our researchers and others have access to public compute through our AI supercomputer. We have current systems that manage that, such as the Digital Research Alliance of Canada, or DRAC, our digital group.

So we have a request for proposal, or RFP, this year, and we will make our selection in terms of how we want to build this public compute, this so-called AI supercomputer, but that is about public access. That ensures public access. I won’t get into the details in terms of why that is important in terms of the size of it, like the petaflops or the power of it, as it were, but it absolutely gives the public, our institutions and others public access to that frontier research. That’s part of it.

There is another element that you’re getting at that I don’t have a lot of time to discuss, which is this: Should we have a public AI? We are looking at things like open systems that exist. We have to make sure those are safe. We have to understand who is building them and how they’re built to make sure that we’re not just giving free access to open large language models that are not safe, or that people’s data is collected.

Thank you, minister.

Welcome. We currently have Bill S-5 in the Senate. It has passed second reading, and it’s on health care systems’ interoperability and is called the connected care for Canadians act. Interoperability, as you know, enables AI to function in areas such as remote patient monitoring, disease prediction and clinical workflow optimization. The improved interoperability within Canada as a publicly funded health care system will provide a firmer foundation for large comprehensive data sets to support research and innovation in health care products and services. What opportunities do you see in Canada’s health care economy, and how should they be encouraged by the government?

Again, I apologize for only speaking English here. It’s harder for me to speak in French, since the ideas are a bit more complex. I want to be as clear as possible, so I will speak in English, but I do understand French, so if you wish to speak to me in French, that’s not a problem.

Health care data is the most important data we have. I want to be clear. Our public health care is a national treasure. It’s one of the great institutions we’ve built, and we have to protect it, and in protecting it, we have to protect the sensitive health care information of our citizens. That data is extremely valuable, and it’s extremely valuable for innovation to serve Canadians better, but how do you do it in a secure way?

First of all, we have to collect the provincial data and try to get some standards to make sure we have access to it all. There are many steps to that, and there is security involved there. The potential to serve Canadians better through access to health care data is important only if it’s done reliably, safely, securely and transparently with Canadians. I’ve worked very closely with Minister Michel, the Minister of Health, who is working with provinces, territories and industry to modernize our health systems and strengthen it for Canadians.

There are countless remarkable health care innovations. I talked about Kardio Diagnostix at the beginning with Dr. Chen. There are many other innovative companies that are using Canadian‑patented wearable nylons that measure heart rates and get diagnostics instantly. We’ve seen that patented in Canada. Making sure that is used carefully, reliably and safely will allow us to serve Canadians better and decrease wait times. Hospitals are experimenting with this, but our responsibility is to make sure that the data is used very carefully, and we’re working closely with organizations like the Canadian Institute for Health Information, or CIHI, and others who know that data space, and I’ll defer a bit to Minister Michel on how that can be done, but it’s an absolutely vital point.

Thank you, minister.

Thank you, minister. This is very last minute.

In December, Canada and the European Union signed agreements to deepen their collaboration on artificial intelligence and digital trust services, particularly through the Canada-EU Digital Partnership Council. I would like to hear your perspective on how these exchanges with the European Union could influence Canada’s approach to AI system standards, governance and oversight.

Thank you again for another great question.

We created an alliance not only with the EU but also with a number of partners to create sovereign options so that we essentially have security and autonomy. Part of that will be through partnerships with trusted countries. This is not just with the EU. I’ve also signed a sovereign technology alliance with Germany, and we’ve worked extremely closely with the Germans on that. We’ve had deepening ties with the Norwegians, who were recently here, and we’ve also signed a technology initiative between India, Canada and Australia.

In general, for the framework, working closely with these countries is vital in a number of areas. First and foremost, we’re making sure that we’re sharing research. We’ve done mapping systems; we’re doing what they call “ecosystem mapping” to see where we can work together on research and where our businesses can work together so they can co-invest. You’re seeing those kinds of co-investments happening all the time. It’s about how we can deepen ties to bolster our sovereign capabilities and increase trade.

The other area, though — and we haven’t mentioned it yet — is safety. AI safety is critical not just on the legislative side, of which we are learning a lot from each other. The Europeans are concerned their safety might have overly constrained innovation, and they’re talking about perhaps rolling it back a bit. We’re talking to them carefully about their rules, like GDPR, and how they function.

There are also technical solutions. Our government has invested with Yoshua Bengio, our A.M. Turing Award winner and the most-cited scientist in the world, on his new frontier AI project called LawZero, which is an AI that is built to police AI systems as a form of technical safety. We have invested in and partnered with them as the Canadian government, and there is massive interest in Europe, in countries like Germany, Norway and others, and in Japan.

So, this is about trusted partners to work together — not only on regulatory solutions for safety and protection but also on technical solutions — as these technologies evolve. This is why Canada has been at the forefront of both.

Senator Deacon, first, I want to thank you for the question and for your ongoing work in this space. You’ve done tremendous work. You and I have had many conversations, and I appreciate them. Many of the senators here have done tremendous work, and Senator Deacon has been right at the forefront of it.

You’ve identified the core problem, senator, which is that we have to move from reliance to resilience. For the last number of years, we have created systems with some of our partners, but we now recognize we have to have options. The reason that we have a “Buy Canadian” policy is not just to support our industry; we need to. It’s not just to create a more robust industry here, but sovereign AI is a form of security and safety.

There are some core examples. One is the renovation of our procurement system, working closely with Minister Lightbound and Mr. Shafqat Ali at the Treasury Board to make sure that not just the supply side, where we give grants or money, but the demand side, which is quicker — you’ve done a lot of work on that — to make sure that the Canadian government procures, through sandboxes and other models, such that we can use Canadian solutions more quickly.

You’ve seen the MOU I’ve done with Cohere. You will see the results of that more specifically very soon as we procure from them, as well as from Coveo and others. By the way, our dual use in defence will also do that.

So, there’s a procurement side that’s key. Our AI Compute Access Fund is also something we’re interested in continuing. It’s $300 million currently. It’s not only to supply compute and support compute access for SMEs, but to make sure it’s designed to support compute access when they use Canadian compute at a higher level, so it has a dual purpose there, senator.

Senator Deacon, that is a great question.

You referenced the sprinkler. Maybe for the benefit of the rest of the chamber, I’ve often said that, in government, we sprinkle money around versus building pools. As a kid, we had a sprinkler, and my neighbour had a pool. No one wanted to come over to our house on a hot day, because a sprinkler doesn’t cool you off. Everyone wanted to play in the pool.

We would like to get people to come to our Canadian pool, and that means we need to build some pools. That means we should be unafraid to champion our champions, and we will be unafraid to champion our champions. Where we have competitive advantages, we will invest in them. You’ll see that in this strategy. We will invest in our Canadian champions and strategic assets. Where we need to partner, we will partner, and when we need to buy — because we may not have all the assets in the entire stack — we will make sure what we’re buying keeps Canadians safe.

Your point is to build the pools, so how are we doing that? Currently, we’ve got the AI Compute Challenge, the $700‑million fund that we’re building. We also have our AI Compute Access Fund. However, I will tell you, senator, very explicitly that the build part — the pillar of build — will be investing in Canadian companies that have the capacity to be world leaders to create those unicorns, but we can’t forget that 95% to 98% of our economy are SMEs. I was a small-business owner for 10 years, and we —

Thank you, minister.

Thank you very much, minister. It is great to see you.

It is good to see you, sir.

You mentioned that, even with all the positives, with social media and AI, there are still a number of real challenges and potential harms, especially for young people. I followed last weekend’s Liberal convention. I guess old habits die hard. I saw two resolutions passed, the first calling for a ban on the use of AI chatbots for anybody under 16 years of age, and the second calling for a ban on social media for those under 16.

In the development of government regulations on platforms for safety and privacy features, as well as “deepfakes,” you spoke about a “pragmatic balance.” Could you share with us what you see as being a pragmatic balance, especially as it applies to children?

First of all, senator, it is great to see you. I appreciate the work you’re doing here and in every single chamber in the precinct.

Senator, protecting children is the core. I’m a father, so I say this personally and as a matter of our mandate as government officials: We have to make sure we protect all citizens, but we have to make sure our children are safe. We’ve seen horrific issues, like Tumbler Ridge. That’s just one case — a tragic case. We know the harms of online behaviour; all of us have seen the impacts it has had on our children, such as loneliness and alienation. This is something I’m discussing very closely with my colleague Marc Miller. Online harms will fall more under his auspices, but we’re talking about AI chatbots and transparency legislation. We will look at that and where that would fall. We’re looking at things like bans.

When you’re making bans, you have to be careful that you’re not doing unanticipated harm to some communities. We’ve heard from rural communities and Indigenous communities that social media is very important because of isolation. I’ve heard this in my riding: The 2SLGBTQIA+ community is saying that people are physically bullied at school, and the only place they find a community of like-minded people has been online, and it has saved their lives. God forbid we take away an online community from a trans child like that who is bullied, and they go back to a physical environment where they’re not comfortable, and there’s self-harm there.

I will just say that we are taking these issues very seriously. There is a legitimate and widespread view that we have to make sure that we protect our children from social media harms, but how do we verify the age, how do we enforce and how do we make sure, by the way, especially when it comes to AI, what the scope of it is? Does it capture video games where you can chat? Does it make sure that our young people have access to technology so that when they’re 16 or 17 — what happens to a generation who might never have been exposed to that? Are they behind kids in Singapore? We have to make sure that we find a right balance.

Minister, since the modernized Official Languages Act guarantees the substantive equality of French and English, what measures is the government implementing to ensure that the AI tools used within the federal government deliver equivalent quality in both languages, without disadvantaging francophones?

This is a very important and critical issue for our country, for the minister and for AI technologies, since the government uses this technology. The Treasury Board has rules in place for this.

They ensure we always comply with the Official Languages Act, that we make sure we have full access for all Canadians who speak French, and we make sure that those systems are available equally in English.

This will also be reflected, again, in our national strategy, but this is not only a core value to our government; we strictly adhere to the Official Languages Act. It is essential; this is not deviating —

Thank you, minister.

Honourable senators, the sitting of the Senate is resumed.