Evidence of meeting #41 for Industry and Technology in the 45th Parliament, 1st session. (The original version is on Parliament’s site, as are the minutes.) The winning word was models.

A recording is available from Parliament.

On the agenda

Members speaking

Before the committee

Sookman  Senior Counsel, McCarthy Tétrault, As an Individual
Reid  Founder and Chief Executive Officer, Mission Control Space Services Inc.
Brian Gallant  Chief Executive Officer, Space Canada
Desrochers  Executive Director, Coalition for the Diversity of Cultural Expressions
Laforce  Executive Director, Copibec
Buteau  Consulting Program Officer, ControlAI
Luccioni  Co-founder and Chief Scientific Officer, Sustainable AI Group

Karim Bardeesy Liberal Taiaiako'n—Parkdale—High Park, ON

Do you have any examples of artists' collectives or cultural industries that are using artificial intelligence to ensure that money flows back to artists and creators in Quebec and other places?

5:20 p.m.

Executive Director, Coalition for the Diversity of Cultural Expressions

Marie-Julie Desrochers

For now, there is no remuneration from developers' use of artificial intelligence. However, our position on authorization and transparency is aimed at a new licencing market.

I'm not sure whether you're referring to technologies like blockchain or to attribution technologies, but things like that are in development. However, for now, because there are no prerequisites, companies do not negotiate with management companies and as a result, there is no remuneration coming down to the artists, unfortunately. Perhaps Mr. Laforce can expand on my answer.

Karim Bardeesy Liberal Taiaiako'n—Parkdale—High Park, ON

Okay, thank you.

I have a few questions for you now, Ms. Luccioni.

I want to start with the demand side and then maybe the supply side of the technologies we're talking about.

On the demand side, you referred to both electricity and water usage. I'm wondering if you can talk to us about some of the innovations you're starting to see on the kinds of water usage that can now be either reduced or recycled for some of the AI data centre applications that we're seeing.

5:20 p.m.

Co-founder and Chief Scientific Officer, Sustainable AI Group

Sasha Luccioni

Essentially, the innovations are often a trade-off between energy and water. For example, if you have an open-loop cooling system, the water evaporates, and it uses more water overall but less energy. If you have closed-loop cooling systems, which are more innovative, you need to use energy for cooling that water so it doesn't evaporate.

Honestly, there's no free lunch, because you're using either one or the other. For sure you can have places where there's more water like a lot of places in Canada. Maybe it's not an issue, but 70% to 80% of it does evaporate.

Karim Bardeesy Liberal Taiaiako'n—Parkdale—High Park, ON

Are there any innovations around capturing waste heat from maybe smaller centres or certain kinds of ways to build AI data centres to capture more waste heat?

5:20 p.m.

Co-founder and Chief Scientific Officer, Sustainable AI Group

Sasha Luccioni

Definitely there are a few. For example, it can be used for heating university campuses and even heating pools. I've heard of salmon farms up in, I think, Norway.

The thing is that these data centres have to be close to the infrastructure. If you're building a hyperscale data centre that's in a rural area far from everything, you can't easily reuse that heat, which is why I definitely think that we should be putting the priority into data centres that are smaller, maybe in basements or integrated into existing infrastructure, because that way you can reuse the heat more meaningfully.

Karim Bardeesy Liberal Taiaiako'n—Parkdale—High Park, ON

Are there innovations in Canada with Canadian cities? We often hear about AI data centre siting as something that's creating real concern. How could some of your ideas help translate into more acceptability in certain locations for AI data centres, if they're scaled appropriately and perhaps more integrated into city planning?

5:20 p.m.

Co-founder and Chief Scientific Officer, Sustainable AI Group

Sasha Luccioni

I think that the data centre that was recently announced in Vancouver is pretty well integrated into the city and into the existing infrastructure. Sadly, the ones that we've heard announced in Alberta and Saskatchewan are definitely bigger and more scaled outside-of-cities approach.

I think we should be really thinking about, instead of building one big data centre, how we build three or five medium-sized ones that are integrated. That way it's easier to leverage renewable energy, because typically it's a lot harder to have enough renewable energy to power a huge gigawatt-sized data centre, where natural gas is easier to burn.

Karim Bardeesy Liberal Taiaiako'n—Parkdale—High Park, ON

As a final question, you've created a company that's helping others reduce their planetary footprint through this technology. Can you talk to us just briefly about the emerging sector of climate AI and other companies that are working in this space?

5:20 p.m.

Co-founder and Chief Scientific Officer, Sustainable AI Group

Sasha Luccioni

The goal is to, first of all, help companies understand the climate footprint of the AI that they use, measuring to help them make more informed choices, for example, and shifting to regions with more renewable energy and greener compute and putting consumer pressure on AI developer companies for transparency. If enough people or organizations that are using AI ask for these numbers, the AI developing companies will have little choice but to provide some form of transparency, which we're currently missing.

The excuse is that people aren't asking for it, so they don't provide it. I think the time has come for the AI industry to be more mature and to push for this transparency, as we have done for other things like transportation, food and things like that.

Karim Bardeesy Liberal Taiaiako'n—Parkdale—High Park, ON

Thank you.

5:25 p.m.

Conservative

The Vice-Chair Conservative Raquel Dancho

You have the floor for six minutes, Mr. Ste‑Marie.

Gabriel Ste-Marie Bloc Joliette—Manawan, QC

Thank you, Madam Chair.

I'd like to start by welcoming our four witnesses. I want to thank them for being here and for their commitment to their respective endeavours.

In particular, I would like to congratulate Ms. Luccioni on her award from the Quebec National Assembly in mid-March.

I'm going to run out of time to ask all my questions, and so I'll get going. I'll start by firing out three questions to Mr. Buteau.

First, can you briefly explain the difference between generative artificial intelligence and artificial superintelligence again? You spoke about Anthropic and what the Mythos model is able to do now compared to other versions of Claude.

Second, in your recommendations, you say that some countries should form a coalition to prevent the development of artificial superintelligence. To your knowledge, have other countries done that and have they made this public?

I'll ask you my third question later, because it has slipped out of my mind.

5:25 p.m.

Consulting Program Officer, ControlAI

Samuel Buteau

Thank you very much.

I'm going to answer in English to make sure I use the correct technical terminology.

To distinguish between superintelligence and generative AI, it's important to understand that there's not a sharp boundary above which we have the risk, where the Nobel Prize winners and CEOs have warned of extinction, and then below which everything is fine. It's more of a continuum.

One of the biggest differences in kind is if the AI is able to autonomously perform tasks on its own for long periods of time, sometimes called an AI agent. The difference between AI agents and generative AI, like chatbots, is often that the AI is smarter being an agent. Then, because it's given more leeway, it has more dangers associated with it. The difference between AI agents and superintelligence is just a difference of degree, and that's why we don't call for unilateral disarmament. It's more a question of having an international trust-but-verify regime where whatever the level of risk is that we deem acceptable, we can go to that front and operate those systems, but there's a way to not go to the point where we actually lose control.

Can you very quickly begin your second question so I can answer?

Gabriel Ste-Marie Bloc Joliette—Manawan, QC

Yes. I was talking about a coalition of countries—

5:25 p.m.

Consulting Program Officer, ControlAI

Samuel Buteau

Okay. I can answer that question in French.

This year, I have personally spoken with more than 120 legislators in Canada. We started in the United Kingdom, where our campaign has now received support from over 110 individuals. We are now working in the United States and Germany. This is one example of what can be accomplished by means other than a country's budgets or diplomacy. I'm just a technical expert that has tried to explain that to their government for less than a year and has made similar efforts in the United Kingdom.

At a deeper level, we find that many people understand the situation once it has been explained. The main problem is that the vast majority of people have not been told about the risks and solutions associated with artificial superintelligence.

Gabriel Ste-Marie Bloc Joliette—Manawan, QC

Thank you very much.

My next question is for the four witnesses. I'm going to ask them to be very brief, if possible, because time is running out.

What do would you like to see in the artificial intelligence strategy that will be unveiled this week?

I know Mr. Laforce touched on that briefly earlier, but I'd like to ask Mr. Buteau, Ms. Desrochers, Mr. Laforce and Ms. Luccioni, in that order, to answer the question as briefly as possible.

5:25 p.m.

Consulting Program Officer, ControlAI

Samuel Buteau

AI is a bit like technology; it's a word that refers to many systems.

There are many concerns that people rightfully have. For example, there is AI adoption, which is very important, and AI sovereignty, which is also very important. This, as we see with Mythos, is also a national security concern.

Does the national security part fit best with the adoption and sovereignty part, or is it better, for example, handled by the Department of National Defence? That's procedural, but I think someone, somewhere, should have their eyes on the national security threat.

Gabriel Ste-Marie Bloc Joliette—Manawan, QC

Thank you very much.

It's your turn now, Ms. Desrochers.

5:30 p.m.

Executive Director, Coalition for the Diversity of Cultural Expressions

Marie-Julie Desrochers

We'd like to see our government follow Australia's lead and issue a statement saying that Canada does not have, and will not have, any exception for text and data mining. This statement must be accompanied by measures requiring transparency around the data used to train these systems.

Gabriel Ste-Marie Bloc Joliette—Manawan, QC

That was short and to the point. Thank you very much.

Mr. Laforce, you're up next.

June 1st, 2026 / 5:30 p.m.

Executive Director, Copibec

Christian Laforce

I would echo Ms. Desrochers's remarks, but I also think it is important that each time the government invests in artificial intelligence companies, the investments must have a clause on compliance with the Copyright Act, because the two go together. It's time to ensure that authors and creators receive appropriate remuneration to ensure they continue to create, as Minister Solomon has said he would like to see.

Gabriel Ste-Marie Bloc Joliette—Manawan, QC

Thank you very much.

It's your turn now, Ms. Luccioni.

5:30 p.m.

Co-founder and Chief Scientific Officer, Sustainable AI Group

Sasha Luccioni

Like the other witnesses, I support transparency and the integration of sustainable development criteria in site selection. We are now positioned to make responsible, sustainable use of artificial intelligence in Canada, but we don't seem to be deploying it in the choices that are being made now. I would like to see it used more.

Gabriel Ste-Marie Bloc Joliette—Manawan, QC

Thank you to the four of you.