PROC Committee Meeting

     I call this meeting to order.

    Welcome to meeting number 30 of the House of Commons Standing Committee on Procedure and House Affairs.

    Pursuant to Standing Order 108(3), the committee is meeting on its study of the current state of civic resilience in Canada and later, in camera, on its study on the challenges regarding special ballot voting.

    Before I continue, I would ask all in-person participants to consult the guidelines written on the cards on the table. These measures are in place to help prevent audio and feedback incidents and to protect the health and safety of all participants, especially our interpreters. There's a QR code for a short awareness video.

    I would like to make a few comments for the benefit of the members.

    All comments should be addressed through the chair. Members in the room, if you wish to speak, raise your hand. Members on Zoom, use the “raise hand” function. We will manage the speaking order as best we can.

    I'd like to welcome our witnesses for today. As individuals, we have Taylor Owen, Beaverbrook chair in media, ethics and communications, McGill University; and Heidi Tworek, director, Centre for the Study of Democratic Institutions, University of British Columbia. We also have, from Equal Voice, Lindsay Brumwell, interim executive director.

    Each witness will have five minutes to deliver their opening remarks.

    We'll go to Professor Tworek for five minutes, please.

     Thank you so much for inviting me to appear before this committee.

    I'm Heidi Tworek, Canada research chair and professor at the University of British Columbia in Vancouver, where I direct the Centre for the Study of Democratic Institutions. We focus on platforms and media, including examining online abuse and harassment, elections, and health communications in Canada and globally.

    While some of our research indicates comparatively high trust in electoral institutions in Canada, I will focus today on two broader challenges to our democracy: online harassment and generative AI.

    First, online harassment of public officials and politicians has become an increasing problem. I began looking at the online abuse of political candidates in the 2019 federal election. Since then, our work has examined many groups of public officials and professionals, including health communicators, journalists and academics. By harassment, I mean intimidation such as death threats or identity-based insults, not rigorous and vigorous democratic debate over important issues that could include, for example, swear words because people feel so strongly.

    I'm sorry to report that harassment is worsening, and its effects are very real. In Quebec, for example, 741 of the province's 8,000 local politicians have resigned since 2021. This is almost 10%, and many of them cited online harassment as a reason. Local politicians were surveyed by Canadian Municipal Barometer in 2025, and 63% had experienced harassment.

    Online abuse and harassment also silence public officials. Any emergency, whether COVID or wildfires, puts officials into the spotlight. Abuse swiftly follows. Some public agencies are now reluctant to put forward spokespeople at all. Without intervention, this can become a vicious spiral of silence, in which institutions and public officials say less because they fear harassment, thus lessening the quality of our public democratic debate and undermining resilience.

    Second, generative AI has supercharged such problems. It lowers the barrier to creating massive volumes of fraud, scams, impersonation, deepfakes and other forms of online harassment. This can happen for political, economic or personal gain, or even for nihilistic violence.

    Canadians' increasing use of GenAI chatbots creates new vulnerabilities. Many GenAI providers privilege U.S.-based information, for example. This is even reflected in some tools designed by Canadian institutions to assess the reliability of claims. For example, when I ask such tools questions about health, they privilege answers from the U.S. Centers for Disease Control over information from the Public Health Agency of Canada.

    A crucial issue here is data voids, which are spaces in which there's little or no high-quality information. A study by the Institute for Strategic Dialogue found that when there were data voids, chatbots in the European Union often reproduced Russian state media, even when those sources were banned. This so-called LLM poisoning enables foreign actors to inject incorrect material that is then reproduced within chatbots without users' knowledge. This is a prime space for foreign interference.

    To address these issues, I offer three recommendations.

    First, the government can use legislation to address chatbots and many of the transparency issues related to the problems I've described, for example, in a revised online harms act.

    For full disclosure, I am a member of the reconvened expert group advising the heritage ministry on online safety, but we, the experts, will have no part in drafting the final bill.

    Second, we need to ensure capacity for long-term monitoring across and between elections, rather than providing short-term grants. I support the establishment of a non-partisan Canadian democracy fund, recommended already by so many witnesses. This could complement British Columbia's proposed centre of excellence for democratic engagement, but any funds need to ensure long-term research too.

    Finally, it remains vital to provide accurate information via trusted channels to the multicultural, diverse population of Canada and to prevent data voids. Public institutions need communications strategies built for the 2020s and the chatbot age. Communications are often the first thing to go in a budget crunch, but the cost of poor communications is a weakened democracy.

    Thank you so much. I look forward to questions.

    Thank you so much.

    We'll now go to Lindsay Brumwell for five minutes, please.

     Good morning, Mr. Chair, vice-chairs and members of the committee. Thank you for the invitation to appear today on behalf of the Equal Voice Foundation.

    We are Canada's only national, multipartisan organization dedicated to strengthening women's participation and leadership in political life. We work across party lines, across the country and at all levels of government to help build a more representative, inclusive and resilient democracy.

    Today's study on civic resilience is both timely and important. At its core, civic resilience is about whether Canadians continue to believe that our democratic institutions are open, responsive and worth participating in. It's about trust, belonging and confidence that individuals can contribute meaningfully to public life.

    One of the clearest ways to strengthen this trust is to ensure that Canadians see themselves reflected in leadership and the decisions they make. According to our national research, Canadians in general do not think politics is very accessible. Less than half believe there are opportunities for Canadians to run or even get involved in politics, and eight out of 10 say that politics are not very welcoming to those who are new to the sector.

    When people believe that politics are only for a narrow group of insiders, participation declines. This is why representation matters—not only as a question of fairness but as a matter of democratic strength.

    At Equal Voice, we often see barriers arise long before someone even considers becoming a candidate or getting involved. Many talented women are interested in serving their communities but hesitate because politics continues to appear inaccessible, adversarial, financially difficult and incompatible with caregiving and professional responsibilities. Others face harassment, online abuse or simply a lack of encouragement or pathways into public life.

    As a result, many capable people self-select before they even reach the starting line. This is a loss not only for individuals but also for Canada as a whole. A resilient democracy depends on drawing leaders from the broadest and strongest possible talent pool. Women are 50% of the talent pool.

     We are also increasingly focused, at Equal Voice, on research and evidence. Strong institutions and good training programs need to be evidence-based and need to use real and evolving information. If Canada wants to improve participation, we already know most of the barriers that exist and who is being left out. We also know there are solutions and actions that are effective in addressing many of these barriers. At Equal Voice, we tackle some of these barriers in the quiet time between elections, when the real work needs to happen. This is where Equal Voice continues to show up.

    We are also known for our multipartisan model, and support for this approach is only growing. We heard this over the almost 18 months of consultations that went into our new strategic plan: lead, connect, compete and govern.

    At a time when many democracies are experiencing polarization, Equal Voice brings women together across political affiliations to build relationships, share experiences and support one another in public life. This unique type of bridge-building is valuable. It reminds us that participation in democratic renewal can rise above partisan divides.

    In our experience and research, there are also areas of hope. There's a great deal of untapped interest in getting involved in politics in a volunteer capacity. In our polling, 58% of women are interested in getting into politics at a municipal level. Young women would be more likely to get involved or run if they knew more about the opportunities and process for involvement.

    As the committee considers recommendations, we would respectfully offer three areas for consideration.

    First, continue supporting public life leadership development initiatives, in the long term, that prepare more Canadians for public life, with both civics training and political literacy training.

    Second, recognize harassment and intimidation, particularly online, as real barriers to democratic participation.

    Third, continue improving research and data collection on representation, participation trends and the pathways into leadership, so future policy is evidence-based.

    Canada's democracy is strong, but strong democracies still require regular reflection. Equal Voice is proud to contribute to this work.

    Thank you. I look forward to your questions.

    That was four minutes and 59 seconds. I appreciate that.

    We'll go to Professor Owen for five minutes, please.

     That's impressive.

    Thank you, Mr. Chair.

    Two days ago, our researchers at the Canadian Digital Media Research Network published a report on a network of 20 YouTube channels—with nearly 40 million views—targeting Albertan audiences over the past year. The channels use AI-generated avatars of Premier Smith and Prime Minister Carney, paid American voice actors reading templated scripts and maps placing the western provinces as the 51st through 54th states of the United States. Real Albertan grievances are being co-opted and amplified by inauthentic infrastructure at scale six months from a provincial referendum on secession.

     Civic resilience in 2026 is not an abstraction. Frontier AI models are advancing on a cadence driven by hundreds of billions of dollars of private investment, concentrated political endorsement across major powers and genuine breakthroughs in capabilities. What follows will almost certainly be a period of rapid social transformation and real upheaval, including labour market shocks, epistemic disruption, reorganized public services and new patterns of political organization and contestation.

     What's more, the institutions we've entrusted with our democratic values are not well suited for this moment. Parliaments, courts, regulators, public broadcasters, universities and the press were designed for slower, more deliberative politics. Their legitimacy rests on judgment exercised in public and held to account over time. This cadence is a feature, not a bug, of our democratic system, but it is being asked to hold against a rate of change it was never built for.

     Trust has eroded over a decade, in part because these institutions were slow to adapt as the information environment was being remade. They enter this transition already weakened and facing a compounding risk. This weakness is not incidental but is too often sought both by foreign adversaries working to accelerate our institutional decline and by a domestic political current that runs from openly anti-democratic movements to a quieter indifference about whether traditional institutions survive.

     Our research has documented three shifts in the information environment that bear directly on the challenge we're about to face.

     The first is social fracturing. We've long understood information fragmentation as something that happens within platforms: filter bubbles on Facebook and echo chambers on X. However, when you look at the information environment as one system, which we now do, by tracking coordinated signals across YouTube, X, Instagram, Facebook, TikTok, Reddit, Substack and all podcasts, what you see is different. Canadians are being fractured as a country. The Alberta network is one example of how the fracture is produced and exploited. When social cohesion erodes at a moment of open hostility from the United States, so does our sovereignty.

     The second is who intermediates information for Canadians. Our work has shown that a majority of political news now reaches Canadians through influencers rather than journalists, and AI systems themselves are now becoming news intermediaries. Our audit of four major AI systems proved that they were trained on Canadian journalism, substituted their own responses for it and rarely credited the newsrooms that produced it. Information integrity was anchored in institutions with editorial standards, libel exposure and professional norms. It's now distributed across individuals and technological systems with none of these yet developed.

     The third is AI-native manipulation at scale. The Alberta network produced 12 times more pro-annexation content than all authentic Albertan YouTube channels combined. The scripts were templated, the voices synthesized or paid for and the avatars fabricated. These are not sophisticated operations, though, because they don't need to be. Generative AI has driven the cost of producing such content to near zero, while platforms continue to carry it in Canada to Canadian audiences at scale. The origin and intent of these operations remain unresolved because YouTube holds the data.

     In my view, then, civic resilience requires two things built together.

     First, we need a regulatory regime that compels platform and AI co-operation with accredited researchers. The legislative vehicles already exist for this: a reintroduced online harms act with strong researcher access provisions, extended to cover consumer-facing AI products.

    Second, we need the independent research infrastructure to make those provisions work. Canada is defunding this infrastructure at the moment it most needs it, as my colleague, Aengus Bridgman, testified to this committee last week. As the Hogue commission's 48 recommendations called for, we urgently need a standing national capacity, at arm's length from platforms and government, with structural funding rather than departmental line items that are vulnerable to budget cycles.

     Wide-ranging efforts are needed to build civic resilience, as this committee has heard, but none of them will be possible if we don't understand and ensure the integrity of the information environment on which our democracy depends.

    Thank you. I look forward to talking about it.

    Thank you so much.

    We'll go to questions. We'll start with Mr. Cooper for six minutes, please.

     Thank you, Mr. Chair.

    I'll begin with Ms. Brumwell. You stated that participation has declined in terms of...what exactly? You were not clear. Could you elaborate on that?

    It's participation from women candidates. Heidi mentioned the number who are leaving government, as well as those who are not engaging and getting more involved or putting their name forward. It's twofold.

    Can you elaborate on the data you have and some of the trends you've identified in that regard?

     Certainly. We have trends showing data from 2017 on and the increase of women who hesitate to get involved and participate by putting their name forward. The trends are continuing to show a lack of interest. It has to do with many of the items that were listed, but we're also finding a lack of support and participation for women who are willing to put their names forward.

     To be clear, has there been an overall decline in the number of women candidates at all levels—federal, provincial and municipal?

     No, it's a mix. Federally, we saw fewer candidates in the last election, but then we were on an upward trend of women candidates. If you look at this at a municipal level, the research and data are spotty. We are seeing differences across the country. In Yukon and B.C., they went past the 50% mark in terms of candidates and elected officials, but in other provinces, we are seeing a decline in participation.

    We also plan on submitting a written submission, and I'll lay out our data and evidence in that to help provide more detail if you need it.

     Sure.

    In terms of where you're seeing a decline, were there upward trends? You alluded to, at the federal level, an upward trend and then a slight decline in 2025.

     Yes. There was across the board—

     You identified similar trends in which a number has moved downward, or are there other factors at play?

     The reasons are some of the ones I listed for the decline, where we've been able to gather them. Yes, in the last election, we had fewer women candidates at the federal level across the board and across all political parties.

    What about in terms of the percentage of women who are getting elected? This has gone up, has it not?

     Federally, it's been static for three elections, at about 30.1% to 30.6%. Even after the by-election, it moved from 30.2% to 30.6%.

     What about at the provincial and municipal levels?

     As I said, it's mixed. B.C. and Yukon are showing highlights and passing 50%, but it's different across all the other provinces.

     Okay.

    Professor Tworek, in your statement, you referenced that, of 8,000 elected officials in the province of Quebec, 741 have resigned since 2021. This sounds like a fairly high number, but how does it compare to previous years? Do you have any comparative data?

     Yes. As Lindsay Brumwell said, the data can be a bit spotty, but as far as we can tell, this is a larger number than usual. It represents another part of this trend.

    Lindsay was talking about the numbers who run, but we're also seeing people not staying in politics as long. There hasn't been a survey of all 741, but when we did a news scan, looking at as many as we could, many of them were talking about online harassment as one reason. We're also seeing a loss of institutional memory, because people don't want to stand for election again or they resign more swiftly than we've generally seen in the past, with online harassment being one of the reasons often mentioned.

    Going back to you, Ms. Brumwell, you cited some of the factors for why there has been a decline in participation, with people self-selecting and not wanting to get involved.

     This is not backed up by data, but my observation is that while there's work to do in Canada in terms of participating in political parties, getting involved and running for elected office, and although there are significant barriers, relative to many other comparable jurisdictions, those barriers are far less significant than perhaps those of the United Kingdom, for example.

    Do you have any thoughts on that?

     I can't speak to the United Kingdom or other countries. What I can encourage the committee to do is to look at all levels of government and women's representation in them.

     The last comprehensive study that was done by the Federation of Canadian Municipalities still showed that 16% of municipal governments do not have a single woman on city council, or municipal council.

     In terms of civics resilience, look at the grassroots, local levels and all levels of government. This shows the more complex barriers and where they still are, versus looking at just a federal level.

     Thank you so much.

    We'll now turn to Ms. Brière for six minutes.

     Thank you, Mr. Chair.

    Welcome to our three witnesses.

    I'll start with Professor Owen.

    In your opening remarks, you mentioned that the institutions to which we entrusted our democratic values are not adapted to the pace of this new information environment. You gave three reasons: social fracturing, AI-native manipulation and the fact that there is an intermediary, which has an effect on integrity.

    I would like to hear you talk a little more about all that.

    A baseline requirement for our democratic institutions to, I increasingly believe, reimagine what their functions are in our democratic society—not to be too bleak about it—requires them to understand the ecosystem in which they operate. They're doing so in a very limited way right now.

    There are primarily research, journalistic and civil society organizations across the country trying very hard to understand a complex ecosystem, when they have limited resources and capacities. The resources are limited because Canada does not have the funding structure for this kind of work that exists in other countries, and operations are limited because the data needed to understand this ecosystem is largely controlled and hidden from view by the companies that have it.

    We on the outside are relegated to a set of unsanctioned research methods, which are regularly being disrupted by companies that do not want us studying this ecosystem, and we have very limited capacity. This is not the baseline status we want if it's even directionally true that our journalistic, legal, university and government institutions are about to go through a fairly rapid transformation, which I personally believe they are.

    Of all the things that need to happen to improve our civic resilience, and there are many.... I know this committee has heard from a wide range of amazing people working in this space and doing incredible things, but I do not know how we operate unless we have a baseline understanding of the environment in which Canadians receive information, while ensuring there's a baseline reliability in it. To me, this is a precondition for everything else, and I don't think we're doing that sufficiently right now.

    It always evolves very quickly, so it's not always easy to keep up.

     I missed the first half of that. I'm sorry, but can you repeat the question? I got the translation and half of the French, but not together.

    I was saying that it always evolves very quickly, so it's not easy to keep up.

    This is, 100%, the underlying variable that has changed instrumentally. The last decade of transformation, as we've shifted from legacy information providers to a new set of actors in our social ecosystem, has been a decade-plus-long transition, and we still haven't kept up with that.

    Now what's happening is the creation of generative AI content, plus a new reliance on AI as an information intermediary and platform itself. Those two phenomena are happening at a radically faster scale than the transition to social media. The problem is amplified.

    You, Professor Owen and Professor Tworek mentioned that the government should get more involved or put in place better things to foster civic resilience, make investments and set up infrastructure for research and data collection.

    Is that correct?

    It is, without question.

    I think two conditions are required.

    The companies themselves have to be mandated to share a certain subset of information that they control and have access to. This has been done in some other jurisdictions, with varying degrees of success, and is one of the core requirements in the duty to act responsibly, which is the main provision of the online harms act. The way it's designed.... The language of the previous version of the bill is quite good and could be adapted to some of the AI issues I talked about.

    Then you need institutions and the capacity to make sense of those data. This is not easy, and it requires collaboration among researchers around the country and a central data steward that is neither the companies nor the government. We do not currently have this infrastructure. It would need to be built if the online harms act were legislated in some version of its previous form—making this a requirement.

    Those two things are the table stakes. Other countries saw this far earlier than we did.

    I'm going to have to interject. Thank you so much.

    We'll now go to Madame Normandin.

    You have the floor for six minutes.

    Thank you, Mr. Chair.

    I'd like to thank all the witnesses for being with us.

    I'll start with you, Professor Owen.

    I understand that, in 2024, you published a study on how Facebook users reacted to the news ban. You noted changes in user behaviour. In particular, screenshots were starting to be used instead of URLs to share information. You noticed, if I'm not mistaken, a decrease in the sharing of disinformation, possibly due to the fact that people back then didn't want to be banned from Facebook and were reluctant to share URL links, even from non-journalistic sources.

    Do you have an update on the behaviour of Facebook users since then?

     We do not have a major update to the study, no, except this: One thing the study showed is that the Meta news.... We're talking about Meta's blocking the ability, not to post, but for Canadians to see the links to URLs directing them to content produced by Canadian journalism organizations on Instagram and Facebook.

    I think you're right to say an important finding is that the primary way it was circumvented was by posting screenshots of news, not posting full links. This opens an interesting question—frankly, for the CRTC, probably not for this committee—of whether this puts Facebook back into the scope of the Online News Act.

    Also, they started relying more and more on content produced by influencers, not journalists. This was the start of a trend that we've seen accelerate very rapidly. We have a recent study showing that there's more news content created and shared by influencers than there is by journalists in the Canadian ecosystem as a whole now. They've surpassed journalism, in fact, by a magnitude of 2:1. There is two times more news content—broadly defined—being produced by influencers than there is by journalistic organizations.

    This was the case in the last federal election too, at a time when you would think content produced by journalistic organizations was at a premium for citizens.

    Another major change that's happened since the study is that we've seen a shift toward distribution of news on other platforms. This was not the case in 2024. We saw an overall decline in total news consumption, across the ecosystem, when Meta blocked it. Now, a lot of it has moved to TikTok. There has been a real spike in news content being performed on and distributed through TikTok.

    Thank you.

    I would now like to ask Professor Tworek and Ms. Brumwell a two-part question.

    Professor Tworek, you mentioned in your opening remarks that there has been an increase in online harassment of politicians. You've also seen an increase in resignations. You mentioned the case of Quebec, where about 10% of politicians had resigned. Unless I'm mistaken, the resignations are mainly on the municipal side.

    Ms. Brumwell, you mentioned that municipal politics are of greater interest to women, but I see a kind of contradiction, because it seems to me that the friction and harassment is mainly on the municipal side because of proximity to the electorate. Municipal politics play rough.

    I'm curious to know what both of you think about that. Why are women more interested in municipal politics, when it seems to me to be the most abrasive place to be in politics?

    I'll jump in. I'm not sure who was supposed to start on the question—my apologies.

    One of the main reasons that women want to get involved at the local level, as we know from our research, tends to come from their family and personal life commitments. Travel to provincial or federal capitals can put quite a strain on any politician. They are looking for ways to be involved in their community associations, their school boards and municipalities. They feel as though they have a better connection to change. These are the reasons we understand.

    We have heard, anecdotally, that a lot of women do not understand the harassment or push-back that can happen at a municipal level until they get into the position. They're not as aware of it, compared to what they hear about in the news at a federal level, for which coverage is much higher.

    This is my understanding of the data we currently have.

    I'll add that it's quite difficult to do comparables in terms of numbers, because there are many more people involved at the municipal level, and as we've said several times, the data is spotty. It's one of the reasons we wanted to work on this level as well.

    What sometimes also happens is that there are young people aspiring to get into politics, and they will work on campaign teams for provincial or federal elections. They will then see the levels of harassment and decide they don't want to potentially continue on a path they may have already chosen. Sometimes, they will then move towards a municipal level instead, or indeed, they will move into different career paths.

    This is another illustration of how the pipeline is broken. Those who are even interested...when they work on campaign teams, they're often the ones tasked with dealing with the harassment in order to shield the candidate. They end up bearing the brunt of it and, therefore, don't want to become involved in the first place.

    Thank you.

     Mr. Calkins, you have five minutes, please.

    Thank you, Chair.

    Ms. Tworek, I'm wondering if you could go back and re-cover the part of your testimony that dealt with generative AI, chatbots, how all that works and how it is, as you said, a prime breeding ground for foreign interference.

    Yes. I'm happy to explain that a bit further.

    As Professor Owen also mentioned, chatbots draw on many, many sources of information. One of them is journalistic and news organizations. In fact, companies such as OpenAI have signed deals with news organizations. Sometimes, when information is reproduced for users who ask a question, sources will be provided so that the user knows this came from The Wall Street Journal, The New York Times or whatever, but often that's not true.

    What we've seen, of course, is that foreign actors seeking to interfere have evolved their tactics. One new tactic that was documented by the Institute for Strategic Dialogue is for Russian state media sources that have been banned in the European Union to produce massive volumes of information. It's super easy to do with generative AI. They produce a lot, and they look for spaces in which there isn't a lot of information already. This means that when chatbot providers are trawling through the Internet for information, they're picking up the only stuff there is on certain topics—which is Russian state media—and then reproducing it. It's basically a back door—

    Is this because the Russian state media looks like legitimate media?

     Yes.

     Okay. This would be the same thing as.... There's Russia Today; we allow, or the CRTC does, lots of things in terms of foreign state media in Canada. Do you have any thoughts or comments on that?

    Yes. Basically, the point of this study was saying that we can debate the merits of whether the European Union should have banned those sources, but chatbot interfaces have become a space for those sources to now become available.

    They're a back door.

    We can absolutely debate whether or not...but this is a new vector that we need to understand.

     Okay.

    Ms. Brumwell, you have a lot of statistics and information in regard to some of the things that keep women from engaging in politics or entering politics. The information would be a lot more meaningful to me if we had a comparator or contrast with why some men don't get involved in politics. The perception is that more men get involved in politics at certain levels, at least federally, but I know a lot of men who would like to have done it but face the same or similar barriers, or they have concerns about getting involved in public life as well.

    Are those reasons any different for a man than a woman?

    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.

    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?

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

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

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

     When I came here in 2006—

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

    We'll turn to Madame Kayabaga for five minutes.

     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.

     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.

    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?

    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.

     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.

    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?

     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.

    Professor Tworek, what do you think?

    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.

    Thank you so much.

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

    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?

    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.

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

    That statistic was only for municipal or local politicians.

    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?

    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.

    Thank you so much, Mr. Cooper.

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

    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.

    Generally, we know that AI impacts women, due to harassment, but unfortunately, I don't have specific data or research. Equal Voice has not undertaken that. I would look to any colleagues or other witnesses to provide clearer, better and more concrete answers on that.

    As I mentioned, we use comparators. In international polling, one of the questions we ask is this: Is there a role for women in political leadership? The fastest-growing group that says straight out, no, is young men under the age of 29.

    I don't know how the programming is going, but I can tell you that, in our research since 2017, out of all the different demographics, this is the group in which we're consistently seeing an increase compared to men in other demographics or age categories. Those men seem to be more welcoming to women in politics.

     That's very disturbing, and it may be an area in which I'd recommend further research.

    Mr. Owen, perhaps you could shed some light on this.

    It's a difficult question, frankly. There's no question that it remains a challenge.

     The information on which the models are being trained—there's no question—have a wide assortment of biases embedded in them. There are zero questions about it. Also, the people building these tools are largely the same demographic that built the previous generation of tools, and we saw some of the implications of that. As well, the final outputs of these products are themselves showing signs of some of these biases. All three of those things are true.

    The AI problem is slightly different, though, when you're referencing social media. This is not based on any technical expertise I have, but there is much more abstraction and distance between the design decisions made by programmers and engineers and the ultimate output of the technological system with AI than there was for social media.

     With social media, there was a very close connection. You would tweak the algorithm for a newsfeed, and this would have a direct effect on how the newsfeed behaved. You could do a content moderation algorithm that had a direct effect on what the user saw and didn't see or what was prioritized and what wasn't.

    With AI, there are two different layers of it.

     One is a model-weighting issue, which people have some control over, but in a very opaque way. We are directing a system. We're not controlling it in the training of a model. It's not entirely clear what the effect of those biases would be in the training of the model.

     The second is on the output of the model and the interface through which we consume it. This is probably where we need to focus more, particularly on audit attention. It's not necessarily auditing language models themselves, but auditing the outputs of the chatbots, which have a layer of design in them and are probably more subject to the kinds of biases you're talking about. I'd probably focus attention on that set of problems.

    Do the others on the panel want to weigh in on that in the last 30 seconds?

     The one thing we see disproportionately affecting women in politics globally is the question of deepfakes. For example, in the U.K. election, we saw deepfakes against female candidates from all political parties, and we know that 95% to 99% of deepfakes are being aimed at women.

    Thank you so much.

    I'd like to thank our witnesses for appearing today and providing their testimony.

    The committee will suspend as we go in camera.

    [Proceedings continue in camera]