Procedure and House Affairs Committee on April 23rd, 2026

[Lindsay Brumwell]

I would say that, looking at the comparative data—we do survey men as well in order to compare with the women in our surveys—the barriers to men getting involved are very similar to those for women, but they have different rankings, different priorities. The top three for women are not feeling qualified, not knowing enough about the process and having concerns about the financial impacts. These impact men, but they're not their top three, at least according to our research.

Harassment continues to be a concern for politicians of all genders. I'm sure all of you on the Hill are very aware of what you're going through. However, we know from the research that it is disproportionately impacting women, as well as visible minorities and other community groups.

I would defer to my colleagues, because we often use much of their research and studies on the impacts of AI and online harassment and how it impacts different groups. Previously, the Samara Centre presented. They have quite a few reports on this as well.

[Blaine Calkins, Conservative, Ponoka—Didsbury, AB]

Thank you.

Ms. Tworek, you talked about the fact that around 10% of politicians have left public office since 2021. You said they have “resigned”. When I hear this word, to me it means they resigned while they were in office. Choosing not to run again means that you're just choosing not to run again. Are you capturing all of this as the same thing? Are you counting anybody who leaves public life on their own, or of their own volition, as resigned?

[Heidi Tworek]

Yes. To be clear, it was both people who chose not to seek re-election and people who left during their period of office.

[Blaine Calkins, Conservative, Ponoka—Didsbury, AB]

Compared to now, that's about, what, six years...? Is that what it would be?

[Heidi Tworek]

Yes. This was conducted before 2026. It was basically four or five years ago.

[Blaine Calkins, Conservative, Ponoka—Didsbury, AB]

When I came here in 2006—

[The Chair, Liberal, Chris Bittle]

You're at five minutes. Thank you so much.

We'll turn to Madame Kayabaga for five minutes.

[Arielle Kayabaga, Liberal, London West, ON]

Thank you, Chair.

I'd also like to welcome our witnesses. Thank you for your testimonies on this really important issue.

Maybe I'd start with Mrs. Tworek, and I apologize if I'm saying your name wrong. I can call you Heidi if you prefer.

You have direct experience in participating in citizens' assemblies. You have seen what the process looks like. From your experience, what makes them work? What tools for rebuilding trust and civic engagement are involved? What would you say are the conditions that have been in place for them to be successful in other places? I basically want to hear your experience in participating in citizens' assemblies and how we can use them as a tool to strengthen civic resilience in Canada.

[Heidi Tworek]

Yes, thank you so much. Professor Owen can also speak to this, as the centre he directs is working on youth assemblies, with gen Z.

There are lots of things to be said for citizens' assemblies. They do many things. We've seen them used very successfully in other countries, such as Ireland. We have international research showing that they are another tool in our tool box. There's nothing that is a panacea, but they are a really useful tool for modelling what it means to have democratic discourse.

I talked about contrasting harassment with democratic discourse, meaning that we want people to feel strongly about issues. It's okay if you say a four-letter swear word. In our research, when we look at the data, we've made the innovation of saying we're happy with that. What we don't want is the harassment. We've seen citizens' assemblies be a space in which people come in with sometimes quite fixed views, confusion, etc. It's a space in which they work through that. There's democratic, vigorous, rigorous discourse that can happen, from which you can end up with really interesting consensuses.

It's a way of modelling what we can do. It shows the value of in-person dialogue and meeting over time. I'd say that all of this goes along with what researchers such as John Forester have written about—what makes for a space that's democratic. We need something with repeat engagement, in which people are a community in some way, where it's bounded and they have a common goal. Citizens' assemblies model all of those things, whereas social media models basically none of them. This is why we end up with such different results.

[Arielle Kayabaga, Liberal, London West, ON]

Mr. Owen, could you also respond and touch a bit on that? Also, beyond funds and dollars, what other institutional, regulatory or cultural levers do you think the government has? Which of those do you think are most underused in the Canadian context, and what more can we do?

[Taylor Owen]

Where Professor Tworek ended is really important. There are multiple ways Canadians can engage with each other in civil society, and we've defaulted to the worst of them for most of our political discourse, which is having our political discussions via the infrastructure and incentives of social media platforms. We all know what this has done to our political discourse and, ultimately, to the strength of our civil society and our civil resilience.

I really believe that there's a massive appetite for these more deliberative forms of civic engagement. Citizens want them, and the transformation that occurs, as Professor Tworek mentioned, is real. We're doing the fifth of our large-scale citizens' assemblies now. This one is with 100 gen Z Canadians, who are represented from across the country, on how they want AI to be developed, as they are the generation that will live through whatever we allow to be developed. They are passionate. They've come up with a set of policy proposals that are meaningful.

You asked what needs to be added to these things to make them work. The challenges are twofold.

The first one is that there is not a meaningful government interlocutor for their findings, because the government chose not to participate actively in it. Its representatives will be attending the formal presentation. It's great that they're doing this, but assemblies have a real risk of bringing people into a democratic process but then disenfranchising them from it if their findings are not meaningfully engaged. I know you've heard from other people at this committee who've argued for some form of institutional capacity to do assemblies, to support them and to engage with them as part of a legislative process. This would be transformative, in my view.

The second is that assemblies only reach a small number of people—the 100 or so who participate in them—at great expense. There need to be ways of expanding this to broader numbers of people. We've been testing different methods, various moments in assemblies, by using digital tools to expand their reach and engagement to tens of thousands of people, as opposed to dozens of people.

Those two things together could be transformative.

[The Chair, Liberal, Chris Bittle]

Thank you so much.

We will now turn to Madame Normandin.

I still owe you from last week, so go ahead for three minutes, please.

[Christine Normandin, Bloc, Saint-Jean, QC]

Thank you, Mr. Chair.

My question is for both Professor Owen and Professor Tworek.

In 2022, the two of you co-authored an article entitled “Platform Governance Needs a Global Response”. You made the distinction between regulating speech, which can be more easily managed by individual countries, and responding to AI, which requires more of a global approach.

Is this way of differentiating the regulation of online speech and the response to artificial intelligence still relevant?

Are there things that individual countries can do on the issue of artificial intelligence, or do we still need a global approach?

[Taylor Owen]

There's a lot there, and I always get a little nervous when something from years ago is.... I don't exactly know what it was.

That being said, the principle still stands that countries have differing national speech laws and that these make a global governance regime for content regulation nearly impossible—and something we should not be aspiring to.

This is the case, frankly, with the outputs of AI-derived speech. What is common across countries, though, is the design of the systems on which they are disseminating and creating speech. I think some common, or at least regulatory, alignment is possible among jurisdictions, and it is probably needed in the face of the behaviour of the American government, which is trying to undermine these efforts globally. Countries are going to need to come into regulatory alignment with one another, at least, on the mechanisms they're regulating.

Both the mechanisms and the objects of those regulations are going to be social media companies, as we previously discussed regarding the online harms act. Other jurisdictions have already built regulatory regimes for this and, increasingly, for consumer-facing AI products, which are themselves platforms through which content is produced and disseminated.

[Christine Normandin, Bloc, Saint-Jean, QC]

Professor Tworek, what do you think?

[Heidi Tworek]

I'm a trained historian, so I can tell you this from a historical perspective: With many other previous technologies, we've seen companies attempt to have global standards for speech, but it has never worked. It's a distinction between.... For companies, it is cheaper to have one standard of content moderation, but in practice and in reality, we have never seen this function, for many of the reasons Professor Owen just mentioned.

It could, in fact, be dangerous because we could default to the authoritarian standard. It's not necessarily the highest standard of freedom of expression we would end up with. There are lots of reasons to end up with somewhat different regulations.

There are many things we can have standardized or aligned. We can think about this in a modular way. Maybe then things will realign with many other countries. We'd do this deliberately because it would be easier to regulate if we had, for example, the same standards of transparency as the European Union. The article you were referring to.... That was us trying to break down, a little, the idea that governance is all about content and that content could have national or regional particularities. Other aspects of regulation—whether it's transparency or data privacy—could be very much more aligned or more global.

That was just a bit of how, if I remember correctly, we were trying to break this down.

[The Chair, Liberal, Chris Bittle]

Thank you so much.

Mr. Cooper, go ahead for five minutes, please.

[Michael Cooper, Conservative, St. Albert—Sturgeon River, AB]

Thank you, Mr. Chair.

Professor Tworek, I'll follow up on your testimony in answer to questions posed by Mr. Calkins relative to what you cited as a decline in the length of service for elected officials.

Can you elaborate on what the metric is based on and where it is—municipal, federal or provincial?

[Heidi Tworek]

Certainly.

This was an example statistic from Quebec. We don't have the same level of granular detail. Perhaps Lindsay Brumwell does for different provinces, in terms of length of service and how this has been changing over time. We've certainly seen it if we simply look at turnover in a place like Quebec. We've seen more and swifter turnover.

This also speaks to something that's worth researching, including with respect to public officials in the spotlight—whether we see more turnover there too.

[Michael Cooper, Conservative, St. Albert—Sturgeon River, AB]

In Quebec, you're looking at municipal and provincial. Were there members of Parliament from Quebec as well?

[Heidi Tworek]

That statistic was only for municipal or local politicians.

[Michael Cooper, Conservative, St. Albert—Sturgeon River, AB]

Okay.

Looking at the House of Commons, what we have seen over the last 40 years has been pretty stable in terms of the length of service, notwithstanding that we have had elections in which there has been significant turnover. I'm thinking of 1984, which saw a massive Tory majority government. The Tories and the NDP were then wiped out in 1993. There was fairly significant turnover between 2004 and the 2011 election, and then a major turnover in 2015.

In fact, I believe 1993 and 2015 had the largest intake of new MPs. Notwithstanding, if you look at the House of Commons between 1993 and 2015, the average length of service was 8.7 years. The average length of service for members of Parliament who decided on their own accord not to seek re-election was 11.2 years. Since 2015, the average length of service has been approximately eight to nine years. There has been no change whatsoever at the federal level. Can you speak to that?

[Heidi Tworek]

I think it speaks to the importance of doing this research at multiple levels.

As I mentioned, I began the research by looking at 2019. To speak to a prior question, we looked at all political candidates for some of the exact reasons you described. We wanted to understand the dynamics of harassment against any individual who is running as a political candidate. I can certainly speak more to that.

We see different dynamics at something like the municipal level, in part because, as Lindsay Brumwell said, people go in not expecting harassment as much. The other reason is that institutions at this level have less capacity to deal with harassment. On the Hill, protections have been stepped up in different ways for lots of reasons. People have larger teams than at the municipal level. The institutions there have less capacity to handle some of the problems they are facing in terms of harassment at in-person meetings and online. This helps to explain the difference at the municipal level versus the federal level.

[The Chair, Liberal, Chris Bittle]

Thank you so much, Mr. Cooper.

We'll now go to Ms. Vandenbeld for five minutes, please.

[Anita Vandenbeld, Liberal, Ottawa West—Nepean, ON]

Thank you, Mr. Chair.

I'd like to go back to the AI and the algorithms.

I'll start with you, Ms. Brumwell.

In the status of women committee, we heard that a lot of people programming these algorithms are young and male. They are, of course, replicating some of their own social biases, including unconscious biases.

Now we have AI. I'm talking specifically about the large language models; they're learning from what is already online, which would already have those biases. I wonder how this impacts women in politics.

I'd like to hear from Mr. Owen, as well, about the kinds of social prejudices and biases that might be pre-existing in some of the algorithms and in the AI chatbots.