Evidence of meeting #52 for Status of Women in the 45th Parliament, 1st session. (The original version is on Parliament’s site, as are the minutes.) The winning word was young.

A recording is available from Parliament.

On the agenda

Members speaking

Before the committee

Dunn  Associate Professor and Director, Law and Technology Institute, Schulich School of Law, Dalhousie University, As an Individual
Parent  Cofounder and Executive Director, Le Centre pour l’intelligence émotionnelle en ligne
Tamil Kendall  Director, Partnership for Women's Health Research Canada
Attard-Frost  Assistant Professor, Department of Political Science, University of Alberta, As an Individual
Sestock  National President, Canadian Federation of University Women
Gillespie  Executive Director, Equal Voice Foundation
Ruddick  Coordinator, Election Tracking and Research, Equal Voice Foundation

Marie-Gabrielle Ménard Liberal Hochelaga—Rosemont-Est, QC

Another risk we face is disengagement from the public space of women and people from LGBTQ communities affected by artificial intelligence.

Certainly, in some respects, the production and sharing of non-consensual images has been criminalized. I have an example in mind: At a high school recently, eight to 10 young people were suspended for producing offensive and degrading images of their teachers. What would you have to say about that?

5:10 p.m.

Cofounder and Executive Director, Le Centre pour l’intelligence émotionnelle en ligne

Emmanuelle Parent

We often hear about horrible situations like that. The thinking is that, if there were no artificial intelligence, there would be no problem, so conversational agents and AI tools are banned for young people.

However, I am going to draw your attention to the fact that, while women and members of LGBTQ+ communities are experiencing the consequences of artificial intelligence a great deal, they are also the ones who benefit the most from it, at least on social media.

When you experience discrimination in the off-line world, when you face homophobia, transphobia or misogyny and you cannot be yourself, the online world may be one of the only places where you have friends, one of the only places where you can go for social support. So, when you want to pass laws or ban something, you have to know that you are not only removing the source of the problems, you are also removing the solution for many of these young people.

Marie-Gabrielle Ménard Liberal Hochelaga—Rosemont-Est, QC

Thank you.

Ms. Dunn, can you talk about the risk associated with these people's disengagement from the public space?

5:15 p.m.

Associate Professor and Director, Law and Technology Institute, Schulich School of Law, Dalhousie University, As an Individual

Suzie Dunn

This is one thing that we're seeing, particularly if you're a woman in leadership and if you're a woman in leadership who talks about feminist issues, reproductive issues—

5:15 p.m.

Conservative

The Chair Conservative Dominique Vien

Thank you, Ms. Dunn.

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

Andréanne Larouche Bloc Shefford, QC

Thank you very much, Madam Chair.

I want to thank the witnesses for joining us today.

Indeed, for this study, as we said, we can no longer ignore artificial intelligence. Now we are looking for solutions. In this committee, we make a particular effort to focus on women's perspectives, of course. But, even though we may have a bias, we can be open to other ideas as well.

Ms. Parent, you finished your remarks very quickly. I think you were going to talk about content moderation. So, first, I will give you the opportunity to finish your opening remarks.

5:15 p.m.

Cofounder and Executive Director, Le Centre pour l’intelligence émotionnelle en ligne

Emmanuelle Parent

Thank you very much.

What I was going to point out is that artificial intelligence is responsible for identifying content on social media that may be harmful and removing it to comply with the community rules published by the platforms. However, we know that this content moderation is deeply flawed, and young women are telling us that. They say that they have already reported a sexist comment, but Instagram told them that it did not break the community rules. Other young people will say that, if you post something online, you have to expect to be insulted, as that is part and parcel of putting yourself out there.

What may be interesting—and I even think it may be a solution to include in Bill C‑34—is to say that safety measures and moderation mechanisms are subject to independent audits to ensure that platforms conduct moderation that complies with the laws. We have to make sure that there are no hateful messages online, so that young people's online safety does not depend on their ability to avoid harm and report incidents. They must realize that, if they post something, they have to be ready to face the consequences. However, the platforms that expose them to such harm must also be held accountable if they fail to moderate the content.

Andréanne Larouche Bloc Shefford, QC

That was my first question for you. I may come back to it later. Thank you.

I will now turn to you, Ms. Dunn. In your opening remarks, you spoke at length about your keen interest in the issue of deepfakes, which have emerged alongside the rise of artificial intelligence. As you explained, this is a phenomenon that has really gained momentum. It leaves lasting scars on the women who fall victim to it. You explained that too.

Bill C‑16, which came into force in July, criminalized deepfakes. Have you had a chance to look at that legislation? What would you like to add to it?

5:15 p.m.

Associate Professor and Director, Law and Technology Institute, Schulich School of Law, Dalhousie University, As an Individual

Suzie Dunn

That feels really interesting, because I think the language around deepfakes is really tricky to get precise. I don't have it in front of me, but the language in that bill is whether it's a reasonable representation of another person. I can't remember what the exact language is.

When we look at deepfakes, it's not as though the harm is that you're creating a perfect replica of someone who is nude or engaged in sexual activity. It's the idea that we should have some control over the way our voices, our faces and our bodies are represented online. One of the concerns I have with the language in that bill is that it might be interpreted as if the image doesn't look entirely real and if the person has a third eye on their forehead or a unicorn horn or something else like this, you'll know that it's not a real person. I think there are some concerns with the language of the act regarding how it will be interpreted by the courts.

We'll have to see, once cases come through, because there's so much you can do with deepfakes, and it's not just that a deepfake looks exactly like a real person. It's the fact that it's a misrepresentation of ourselves. I think much of the focus has been on sexual deepfakes, but this goes beyond sexual deepfakes. Think about the harms that occur to people when their images and their faces are co-opted. We see this amongst politicians who have had deepfakes where they're saying things that they didn't say. It's causing confusion and disruption. There has been this focus on sexual deepfakes, but we might need to think more broadly about what rights we have to our images and our voices.

I think there is a gap in the law regarding creation. I'm not sure if criminalization is the perfect answer, but if you look at children's content, we criminalize the creation, the access and the possession—all of these things. With sexual deepfakes, it's still permissible to create that content. I think there are some questions about that. If we're not actually stopping it at the creation point, the only thing we're dealing with is post-distribution, when the harm has already occurred. I think there are interesting legal questions around what the legalities are around creation. Do we need to think beyond just sexual deepfakes, then, and ask how we can protect people from their faces and their voices being co-opted more broadly?

Andréanne Larouche Bloc Shefford, QC

Ms. Parent, we spoke with you about the impact online written publications can have on young people.

Ms. Dunn, we talked about the impact of deepfakes going forward. So we can see that there may be negative repercussions.

I will now turn to you, Ms. Kendall. We have less than a minute.

According to a 2025 Canadian Women's Foundation report, 61% of women and gender-diverse people have experienced cyber-violence because of their gender identity. For all these reasons, you talked a lot about research. We could come back to that.

When it comes to mental health, what—

5:20 p.m.

Conservative

The Chair Conservative Dominique Vien

I am sorry, Ms. Larouche, but you will have to continue later. Your time is up.

Good afternoon, Mrs. Borrelli. You have the floor for five minutes.

5:20 p.m.

Conservative

Kathy Borrelli Conservative Windsor—Tecumseh—Lakeshore, ON

Thank you, Chair.

I'd like to welcome and thank all of our witnesses today. My first question is for Dr. Kendall.

How can AI digital tools help women and gender-diverse individuals access reliable medical information and advice, particularly when access to health care professionals and specialized services is limited?

5:20 p.m.

Director, Partnership for Women's Health Research Canada

Dr. Tamil Kendall

That's a great question. I think the answer may also respond to Ms. Larouche's question.

We need to put guardrails around the chatbots so that they are using reliable medical information and, as I also tried to argue, that is not old medical textbooks. We know those are biased. We need to put out the newest research that is actually true to the experiences, and we need more of that research.

We also need them to appropriately refer. That has to be a guardrail. You may have seen the article in The New York Times about the woman with epilepsy who died because the chatbot encouraged her to go off her medication. We're seeing this, and there needs to be regulation from government. I think this is what my colleagues are saying as well: There need to be audits and regulations.

Thank you very much.

5:20 p.m.

Conservative

Kathy Borrelli Conservative Windsor—Tecumseh—Lakeshore, ON

Is there any Canadian data available on the number or proportion of women using AI tools for medical information or advice?

5:20 p.m.

Director, Partnership for Women's Health Research Canada

Dr. Tamil Kendall

That's a great question. I do not have that answer for you, Ms. Borrelli. I will find the answer and come back to you on it.

5:20 p.m.

Conservative

Kathy Borrelli Conservative Windsor—Tecumseh—Lakeshore, ON

Thanks so very much.

Dr. Kendall, how can Canada remain responsible with the usage of AI in health while protecting the consent and confidentiality of the client?

5:20 p.m.

Director, Partnership for Women's Health Research Canada

Dr. Tamil Kendall

This is a very challenging area, as we are learning in the field of health research and also in clinical delivery of care. Privacy regulation needs to evolve as AI is evolving. In the research space, the safest way is to create safe research spaces—like lockdown areas where the data is—and to vet the researchers who will access them.

Again, it's evolving so fast. As legislators, there's a real need to evolve with it, so congratulations to the committee for studying this.

5:20 p.m.

Conservative

Kathy Borrelli Conservative Windsor—Tecumseh—Lakeshore, ON

Thank you.

Ms. Dunn, in the Government of Canada report “Canada’s National Artificial Intelligence Strategy: AI for All”, where do you see gaps between the government strategy and the concerns surrounding AI and gender-based violence that we are discussing today?

5:20 p.m.

Associate Professor and Director, Law and Technology Institute, Schulich School of Law, Dalhousie University, As an Individual

Suzie Dunn

There's a challenge, globally, in this desire for countries to really embrace AI and have us learn how to use it properly. I deal with this with my law students all the time. We know that AI is going to be a legal tool that they're going to have to use, but what are the ways that we can teach them to use it properly and ethically, without it hallucinating cases and all of that? A challenge in “AI for all” is the push forward toward embracing AI in all circumstances. We don't need AI in all circumstances. We don't need AI for “all”. If we are going to have this message, if it is the government policy that we're going to have AI for all, I would advocate for safe AI for all and accurate AI for all.

When we're thinking about creating policies on how we regulate AI and how we embed safety practices and privacy practices—especially when we're going to look at government making decisions using AI tools, which will impact the rights of people—what are the proper procedures to have in there to ensure that we're not going to have discriminatory outputs and to ensure that we have people who have deep consideration for values like equality, dignity and autonomy engaging in those processes?

Rather than saying that AI is good for all things so we should put it in schools, doctor's offices and everywhere, we need to question whether we actually need AI in those spaces. If we are going to put AI in those spaces, we need to ensure that it is safe. If it's not safe, we should not say, “AI for all”. We should say, “No AI.”

There's a bit of a challenge in the government messaging saying that it's AI for all and also trying to recognize the risks of AI.

5:25 p.m.

Conservative

Kathy Borrelli Conservative Windsor—Tecumseh—Lakeshore, ON

Thanks so very much.

I believe my time is just about up.

5:25 p.m.

Conservative

The Chair Conservative Dominique Vien

You are absolutely right, Mrs. Borrelli. Your time is up.

Ms. Gladu, the floor is yours for five minutes.

Marilyn Gladu Liberal Sarnia—Lambton—Bkejwanong, ON

Thank you, Madam Chair.

Thank you to all of the witnesses for being here today.

We've heard testimony already about how, with the people who are creating AI, there's already a gender bias happening. I'm concerned about that, and I'm concerned that we don't actually have the disaggregated data that we need in order to start building models that are not gender-biased.

In addition to that, although Bill C-34 creates a digital safety commission that is basically an independent regulator that is going to ensure that all the platforms have the right things in place to prevent harms, what about disinformation?

I look at things and I don't know whether they're AI or not. I don't know if they're true or not, and I've seen things that look like me saying stuff that I didn't say. What are we going to do about that?

Let's start with you, Ms. Dunn.

5:25 p.m.

Associate Professor and Director, Law and Technology Institute, Schulich School of Law, Dalhousie University, As an Individual

Suzie Dunn

That's the hardest question. Misinformation and disinformation are very difficult.

One area to regulate is bots. We know a lot of bot content is producing a lot of misinformation and disinformation. In those circumstances, it's about trying to create worlds where at least it's just humans who are creating that content. For misinformation and disinformation that's discriminatory, we should be able to regulate that on platforms to say that it is inappropriate behaviour.

Who interprets what is misinformation, what is disinformation, what is art, what is satire and what is just people being wrong? There is an idea that a central truth exists in some areas, but often it does not.

I think we have real risks to democracy and to truth. I've heard stories of AI videos being posted, and then someone says, “That's an AI video. It wasn't you who said that.” The person who posted it then says, “It doesn't matter. She probably would have said that.” As long as it suits what their beliefs are, AI is used as a tool. We see politicians creating AI of themselves to represent themselves in a particular manner.

The question of AI misinformation and disinformation is one of the messiest ones when it comes to legal regulation. I don't have a great answer beyond that, but it's a major risk.

Marilyn Gladu Liberal Sarnia—Lambton—Bkejwanong, ON

No, that was very good.

Ms. Parent.

5:25 p.m.

Cofounder and Executive Director, Le Centre pour l’intelligence émotionnelle en ligne

Emmanuelle Parent

I do not have the perfect solution. Obviously, that is what we are here for.

When we tell young people to check their sources, we like that, do we not? It makes us feel smart to tell them to check their sources, to look at their sources. However, when it is 11 p.m., you are on TikTok and you are watching your 23rd video, do you check your sources? That is not necessarily the case. Young people are not going to check all their sources either.

However, we see a general sense of skepticism, which can be a good thing for critical thinking. For example, when we share a piece of information with people, we say we saw it on TikTok. Doing so encourages others to think critically as well.

One thing that can help is having spaces where we know the information to be of high quality. This is not my area of expertise, but anything that aims to fund high-quality, impartial information that prioritizes the collective good is a good solution. This is because telling young people to check their sources on TikTok is not realistic, but telling them to visit the Radio-Canada website to find out what is being said there makes sense.