Information & Ethics Committee on Feb. 5th, 2026

[The Chair, Conservative, John Brassard]

I'm going to call the meeting to order.

Welcome to meeting number 26 of the House of Commons Standing Committee on Access to Information, Privacy and Ethics.

Pursuant to Standing Order 108(3)(h) and the motion adopted on Wednesday, September 17, 2025, the committee is resuming its study of the challenges posed by artificial intelligence and its regulation.

With us today we have the Honourable Evan Solomon, who is the Minister of Artificial Intelligence and Digital Innovation, and with him from the Department of Industry is Mark Schaan, who is the associate deputy minister.

Minister, I'm going to allow up to five minutes to address the committee and then we'll be following that with questions.

Thank you.

[Evan Solomon, Liberal, Minister of Artificial Intelligence and Digital Innovation]

Thank you, Mr. Chair.

Members of the committee, thank you so much for the opportunity to appear before you today.

As you said, Mr. Chair, I am joined by officials from my department. I appreciate the committee's deep interest in this kind of work.

As some of you know, I've already appeared before the science and research committee in the fall session and again just this week to discuss my mandate and share with committee members the work we're doing to build trust with Canadians concerning artificial intelligence.

I am honoured to be here today and to participate in this essential study on artificial intelligence.

To put it simply, AI is a transformative technology, and if used properly, it will improve people's lives and serve Canadians for the better. That is our mission, and that is our goal. I'm glad this committee is here to dig into the foundational issue we need in order to have responsible, reliable AI, and that is trust.

If Canadians want to reap the full benefits of artificial intelligence, they must be confident that this technology is being used safely, fairly and responsibly.

The countries that succeed will be those that use this technology responsibly in ways people can trust and that actually make life better. AI is meant to serve people, not the other way around.

As the first Minister of Artificial Intelligence and Digital Innovation for Canada, I've been working towards a renewed national AI strategy for Canadians almost two years ahead of schedule. Trust, safety and responsibility are the foundations of that work.

Our AI strategy is anchored in one core principle and three pillars.

The core principle is very simple; it is “AI for all”, and the three pillars are build, empower and protect.

“AI for all” means this: No matter where you live in Canada and no matter your background, your age or your income, this technology will work for you responsibly, reliably and safely. It will strengthen our democracy and deliver better public services if we manage it correctly. It will create good jobs for Canadians and protect people, especially children and vulnerable communities, from harm. That's what the principle is.

“AI for all” is already being put into action. It is guiding the final stages of our refreshed national AI strategy, which we'll be launching very soon. Our strategy is shaped by real input, ideas shared by Canadians through our public portal, expert advice from our AI task force—their reports are now available to the public online—and insights from researchers, workers and industry leaders across the country.

The goal is simple: to set a clear direction where Canada is going in AI and how we will get there on our own sovereign terms for our benefit.

If I may, I'll walk you very briefly through what this means in practice and our three pillars: build, empower and protect.

With respect to the first pillar, we are building a strong, sovereign and safe AI foundation to drive economic growth and create prosperity. That requires the infrastructure and, in plain terms, the compute power that allows Canadian companies to build and use Canadian technologies right here at home.

That means security for Canadian data and peace of mind for the citizens of Canada. That means more opportunity for Canadian innovators, more resilience for our economy and more jobs staying and being built right here at home. This also means digital sovereignty. Sovereignty does not mean isolation—we're a trading nation—but it does mean having the capacity to choose where we build and where we scale to keep Canada under Canadian law.

It is a matter of digital sovereignty and choice for Canada.

The second pillar is empower. Canada already has extraordinary strengths in talent and research. I just look at our three great AI national institutes: Mila in Montreal; Vector in Toronto; and AMII in Edmonton. They're global anchors of excellence, built by pioneers who helped shape modern AI.

Our focus now is to shorten the distance between impact and insight, turning Canada's world-leading AI science into real-life adoption, higher productivity and companies that scale here at home.

I'll give you an example. We have a new initiative in just the last month: the $100-million venture scientist fund launched by Mila and the VC company Inovia to invest in scientists to turn their ideas into action.

That progress only matters if Canadians are part of it. Investing in skills and training is an essential part to make sure people in every region succeed in our AI-driven economy.

The third pillar is protect. Trust must underpin everything we do. It is the foundation. That's why in 2024 we launched the Canadian AI safety institute, bringing together leading researchers from inside and outside government to focus on the real risks and the real safeguards associated with advanced AI. This work is connected with international partners, because safety in AI is a global challenge that requires global co-operation. We're committed to modernizing Canada's privacy framework to protect Canadians' data, safeguard children online and ensure as AI evolves that people's rights remain at the centre of our approach.

Protection means that Canadians will have the tools to protect them from harm.

AI is moving quickly. So are we. We're building in Canada. We're buying in Canada. We're building a safe AI for all Canadians to thrive in the future.

Thank you.

[The Chair, Conservative, John Brassard]

Thank you.

We're going to questions.

For the first round of six minutes, Mr. Barrett, go ahead.

[Michael Barrett, Conservative, Leeds—Grenville—Thousand Islands—Rideau Lakes, ON]

Minister, I have a bunch of questions. They're going to be brief. I appreciate that it's a complex topic, but I'd ask that, if you could, you keep your responses equally brief.

The first thing I'd like to touch on with you is that the Canada Revenue Agency developed Charlie the chatbot. Charlie the chatbot did a really bad job. Its reliability certainly did not meet minimum service standards or the expectations of Canadians who have to get it right to the penny when they deal with CRA.

The chatbot didn't even get it right half the time with the advice it gave. Are you drawing from this lessons that you are going to apply to the use of AI in its rollout across the federal public service or the government?

[Evan Solomon, Liberal, Toronto Centre, ON]

Yes. I appreciate the question.

I'm obviously aware of how our government is scaling the use of AI across departments and agencies. There is an effort to drive transformation across government and support a more efficient public service so that we can serve Canadians better and, of course, AI is meant to serve public servants. We have CANChat and GCtranslate.

I will say that the Treasury Board Secretariat has a public register of hundreds of government uses of AI. When something is not functioning as it should, obviously there are specific measures to make sure that action is taken to serve Canadians more accurately and in a better fashion—absolutely.

[Michael Barrett, Conservative, Leeds—Grenville—Thousand Islands—Rideau Lakes, ON]

Minister, just for clarity, this cost taxpayers $18 million and 66% of the time it provided incorrect information to tax filers. Is that acceptable or unacceptable?

[Evan Solomon, Liberal, Toronto Centre, ON]

I think that in every moment when the government serves Canadians, Canadians demand the highest standards and the highest form of efficiency. I know that the Treasury Board and PSPC have rigorous regulations and rules about how they procure, and the standards, and absolutely, meeting the high standards is the expectation of Canadians in any department. Working to improve that is core.

[Michael Barrett, Conservative, Leeds—Grenville—Thousand Islands—Rideau Lakes, ON]

Minister, you'd say that 34% accuracy is not acceptable.

[Evan Solomon, Liberal, Toronto Centre, ON]

I would say that the goal for the use of any technology is to make sure that we serve Canadians with the highest accuracy level and the highest efficiency.

[Michael Barrett, Conservative, Leeds—Grenville—Thousand Islands—Rideau Lakes, ON]

I wouldn't say that was the most efficient way to say that getting the answer right 34% of the time is unacceptable. You can have the opportunity to say that in your next response.

We've seen that the government is going to make a move to reduce access to information for Canadians by limiting the retention of instant messages and email messages. We saw in the ArriveCAN scandal that key actors simply just deleted files. We've seen that departments are taking years to respond to access to information requests.

What commitment or guarantee are you prepared to make that all of the audit trails in fact will be part of the integration and rollout of any AI systems in the government? Are you prepared to make that commitment today?

[Evan Solomon, Liberal, Toronto Centre, ON]

It's a great question.

I want to say two things about how both the Treasury Board and Minister Shafqat function under this. All procurement and all the regulations are followed rigorously to make sure there's transparency and accountability in our communication and in how any procurement and any use of this technology will function.

[Michael Barrett, Conservative, Leeds—Grenville—Thousand Islands—Rideau Lakes, ON]

As an example, when an email is sent, there should be the opportunity for access to information requests to be made, but that happens from person to person. If there's simply work product being generated through an AI system, there's no paper trail of how we got to the end result. If the end result is simply printed out on a piece of paper at an office and the audit trail disappears instantly or after 15 days, obviously that doesn't provide the level of transparency that Canadians would look to have by getting the information, for example, through an ATIP, through an inquiry of ministry or by asking individuals to appear before parliamentary committees.

We don't have that option with AI, unless you would be personally responsible for coming to see us every time we had a question about an AI system.

[Evan Solomon, Liberal, Toronto Centre, ON]

I appreciate the question.

To make sure that we're clear, all records of business value are retained. Obviously, records that are transitory are not retained. They're transitory in nature.

Again, we follow all of the rules in terms of our record-keeping. When we're making a decision about procurement, as Treasury Board will, we're following the guiding principles across all of that.

I will say that Treasury Board work does have some principles on the automated decision-making system and guides for generative AI usage. There are guidelines for that.

[Michael Barrett, Conservative, Leeds—Grenville—Thousand Islands—Rideau Lakes, ON]

Minister, can we expect new rules and public consultations on what that's going to look like with respect to deepfakes and how we're going to protect children from the use of sexually exploitive and non-consensual image creation through AI systems?

[The Chair, Conservative, John Brassard]

I need a very quick response.

[Evan Solomon, Liberal, Toronto Centre, ON]

It's an excellent question.

Absolutely. We already have a bill, tabled by the justice minister, that it would make it a criminal offence to share sexually explicit...that deepfake imagery. We would love to have support on that.

I have lots more to say about deepfakes. It's a key question.

[The Chair, Conservative, John Brassard]

Thank you, Minister.

Ms. Lapointe, you have the floor for six minutes.

[Linda Lapointe, Liberal, Rivière-des-Mille-Îles, QC]

Thank you very much.

I would like to welcome the witnesses and thank them very much for being here today.

I believe you were referring to Bill C‑16, the bill introduced by the Minister of Justice, which will deal precisely with what my colleague mentioned earlier.

You said earlier that a strategy on artificial intelligence would be launched soon. Can you be a bit more specific?

[Evan Solomon, Liberal, Toronto Centre, ON]

Again, I'm happy to talk about the bill we've put forward on the sharing of non-consensual sexualized imagery, especially those generated by deepfakes, the so-called synthetic imagery. We consider that a form of a criminal offence. We have legislation about that. This is a form of violence against vulnerable communities that we will be taking action on in some legislation. We've tabled some legislation and there will be a suite of legislation to deal with that part of our privacy, updating our privacy laws, which is part of my remit, but also I know part of Heritage under Mr. Miller.

These are genuine concerns Canadians have. In terms of our strategy, maybe it's good to just tell Canadians that we had a national task force of 28 different experts from across the country who submitted their reports from all different aspects, some on safety, some on workforce, some on skills and training, some on infrastructure, some on adoption. They're very thorough. They're very interesting.

We had a series of round tables. The public can see all those. It's very transparent.

We also had more than 11,000 public responses, which is the largest, I think, in the ministry's history. Those were long, thoughtful responses from Canadians. They're posted as well, so people can read them.

We've also had multiple round tables. This morning I had a fantastic round table. It is Black History Month, and I had a round table with black Canadians in the tech space who had real concerns about things like bias. We talked about transparency and automated decision systems to make sure that there's not built-in bias. Those are the kinds of things that we're looking at in our legislation.

In our AI national strategy, as I say, building the infrastructure is part of it. Empowering a workforce, skills and training, to make sure the workforce of the future is here, are critical. But for this particular committee, I know you're seized with the protect side, the ethics of it. This is a really important conversation. How do we have responsible, reliable AI to get the benefits and mitigate against the harms? Part of that protect strategy is our legislation, and part of that protect strategy is to build sovereign data centres and sovereign companies because we believe sovereignty is a form of safety.

[Linda Lapointe, Liberal, Rivière-des-Mille-Îles, QC]

Thank you very much. We still have some time left.

I found it interesting that you brought up deepfakes and non-consensual images again. Everyone is definitely interested in that. You talked about the pillars. You talked a little bit about your strategy. You have held several consultations and are making progress on the issue. I assume you will analyze all of this before presenting the strategy.

You also mentioned the three pillars you are working on. Could you give us more details about these three pillars?

[Evan Solomon, Liberal, Toronto Centre, ON]

Sure.

Thank you for giving me the opportunity to talk about them.

Let me just dig, maybe I can get a little more detail on them.

Just so Canadians are aware, we have something called the Canadian sovereign AI compute strategy, which is a $2-billion investment so that Canadian businesses, innovators and researchers have access to the compute capacity. Some people think compute capacity is kind of like the fuel that you need to fuel innovation and ingenuity. That's really important. We've got our AI compute access fund where we have targeted companies to help support getting access to the compute they need for innovation.

Ninety-five per cent of our economy is small businesses, and they need access to the tools or they can't become productive. That's helping. Our compute challenge is also helping build infrastructure.

We're supporting researchers as well. This year we will be building a supercomputer that will be available for universities and researchers across the country. It will make us in the top class of the G7, and will allow us to remain at the forefront and the frontier of research. There's a research side, there's an infrastructure side, there's support for businesses, and again, there will be an aspect of training and skills because that's really important. AI literacy, making sure that people are comfortable and understand this technology and have the skills to use it, is critical.

[Linda Lapointe, Liberal, Rivière-des-Mille-Îles, QC]

Thank you very much. Your comments are interesting.

You mentioned an investment that would benefit universities, if I understood correctly.

Do you have sufficient funds to invest? You mentioned a super data centre. Can you elaborate on that?

[The Chair, Conservative, John Brassard]

I need a response in less than 10 seconds, Minister. I'm sorry.

[Evan Solomon, Liberal, Toronto Centre, ON]

We have the AI compute challenge. It's a $700-million fund. We've already funded that. Yes, we absolutely have the horsepower to get this up and running.

[Linda Lapointe, Liberal, Rivière-des-Mille-Îles, QC]

Thank you very much. That was very interesting.