Survey of Online Harms in Canada

A relatively small number of technology companies have succeeded in largely consolidating and privatizing the online public square.¹ Large online platforms are increasingly used by people in Canada to connect with friends and family, stay up-to-date with the news, and engage in civic and democratic discourse. The impacts of this new public square are continuing to evolve and still not fully understood. However, evidence has mounted over recent years about its negative effects on Canadians’ safety, social cohesion and democracy: hate speech and harassment that target marginalized groups are on the rise; disinformation, both foreign and domestic, is fueling radicalization and extremism; and real-world violence, including sexual abuse and child exploitation, is unfortunately an increasing reality.^{2, 3, 4, 5, 6}

As a result, there have been growing calls for public policy changes to mitigate these harms and rebuild our public square.^{7, 8, 9} At the same time, legitimate concerns have been raised regarding censorship — and that any changes may unreasonably limit our rights and freedoms, particularly the right to free expression.^{10, 11}

In September 2021, we released Rebuilding Canada’s Public Square, which provided the results from our past three surveys on this topic, as well as advice on how to improve the Government of Canada’s proposal at the time to tackle some categories of online harm. Since then, the federal government established an expert advisory group on online safety, which provided advice on how to design a stronger legislative and regulatory framework to address harmful content online. The group rightly suggested a move away from a monitoring and 24-hour takedown model for harmful content, which could drive over-censorship, toward more of a systems approach that places duties on platforms to act responsibly, increase transparency and mitigate their systemic risks.

Public policy on online safety should be informed by evidence about Canadians’ experience with online harms, as well as Canadians’ views on the appropriate role of government in addressing those harms. This report is intended to provide up-to-date evidence about online harms and Canadians’ views on how to govern online platforms — and provide advice on how to do so in a manner that protects and advances Canadians’ fundamental rights and freedoms.

¹ Nyabola, N., Owen, T. & Tworek, H., 2022

² Canadian Race Relations Foundation, 2021

³ Garneau, K. & Zossou, C. , 2021

⁴ Bridgman, A.. et al., 2022

⁵ Humphreys, A., 2020

⁶ Statistics Canada, 2022

⁷ Housefather, A., 2019

⁸ Public Policy Forum, 2022

⁹ Tusikov, N.,2022

¹⁰ Carmichael, D. & Laidlaw, E., 2021

¹¹ Canadian Heritage, 2022

Canadians’ Experiences on Online Platforms

Overall Use of Platforms

Use of online platforms among people in Canada remains high and is growing, with 98% using at least one major platform. While YouTube remains the most widely used platform, Meta’s four major platforms (Facebook, Messenger, Instagram and WhatsApp) have all increased in overall use since 2019 (Figure 1). TikTok is continuing to experience the fastest rate of growth among Canadians, nearly tripling its reach from 10% to 29% in two years, and is now used by more than half of those aged 16-29 (Figure 2). Platforms with noteworthy decreases in overall use since 2019 include Pinterest (-7 percentage points) and Twitter (-6 percentage points).

Platforms As a News Source

While legacy media from television (58%), news websites (44%) and radio (39%) continue to be the top news sources for people in Canada overall, each experienced declines in 2022 from surges in use during the pandemic (Figure 3). In contrast, the use of online platforms as a source of news continues to grow, with 32% citing Facebook, 24% using YouTube and 16% naming Instagram.

Younger Canadians, in particular, are using online platforms for news — those aged 16-29 use Instagram (44%), Facebook (36%) and YouTube (36%) for news at greater or comparable rates compared to legacy media, such as TV (43%), news websites (34%) and radio (27%). The use of Instagram for news among younger Canadians has increased in particular from 19% in 2019 (+25) compared to Facebook (+4) and YouTube (+8). TikTok is also an emerging news source, with 21% of those aged 16-29 citing the platform, growing to 30% among those aged 16-23.

Those using Facebook, YouTube or Instagram for news are more likely to also use each of those platforms for news, and are less likely to use legacy media (TV, radio and news websites).

In addition to consuming news on social media, a significant proportion of Canadian residents actively engage with news and politics on these platforms. In this survey, 35% of respondents indicated that they have commented on or posted links, videos or images about news or politics; 17% have joined an online group about an issue or cause with people they didn’t know; and 13% said they have participated in a government or political consultation or engagement online.

Trust in Platforms

Canadians’ trust levels in online platforms continues to decline. When asked to assess their trust in various organizations to act in the best interest of the public on a scale of 1-9 (with 1 being the lowest level of trust and 9 being the highest), TikTok, Facebook and Twitter had the lowest levels of trust (Figure 5).

This finding is consistent with our past surveys, where social media platforms had lower levels of trust than oil companies, telecommunication providers and mainstream media. Trust levels in social media platforms declined significantly from 2021, with just one in ten Canadians now having high trust (Figure 6). Trust in TikTok fell in particular, replacing Facebook as the least trusted organization.

In comparison, Canadians’ trust in mainstream media outlets remained relatively high, with just one in eight having low trust in CBC, CTV, Global News and The Globe and Mail. Consistent with global trends,¹² trust levels in Canadian media have fallen slightly (Figure 7), and it is worth noting that the decline is concentrated among those on the right end of the political spectrum (7-9 on a 9-point scale). Low trust in these three media outlets has approximately doubled since 2019 among the 16% who identify on the right, from an average of 13% to 25%, while remaining stable among those on the left and in the centre at 10%.

¹² Newman, N., et al, 2022

Exposure to Online Harms

Canadians reported frequent and increasing exposure to harmful content, though self-reported rates of exposure on online platforms have not increased significantly in recent years.

We first asked respondents if they think there has been change in the amount harmful content, such as hate speech, harassment and false information, that Canadians have been exposed to over the past few years. Just over 70% thought that Canadians have been exposed to more harmful content, with only 6% thinking it has declined (Figure 8). We then asked how frequently they see a range of harmful content on online platforms, such as false information, hate speech, identity fraud and promotion of violence (Figure 9).

Survey respondents were also asked how much truth they thought there was to eight statements of misinformation on a range of topics, including COVID-19, climate change, immigration and the Russian invasion of Ukraine (see Methodology section for more detail). A majority of Canadians (53%) correctly identified at least 75% of the misinformation statements (i.e., 6-8 statements correctly identified, categorized as low belief in misinformation) (see Figure 10).

In total, 15% of respondents correctly identified 25% or fewer of the false statements (i.e., 0-2 statements correctly answered, categorized as high belief in misinformation). This group was more likely to identify on the right end of the political spectrum (7-9 on a 1-9 scale; correlation coefficient of 0.5). Weak correlations were also observed with lower household income (0.1) and education levels (0.1), with no significant differences in belief in misinformation by gender or overall frequency of social media use.

High believers in misinformation were significantly less likely to trust authoritative information sources to act in the best interest of the public, reporting three times the level of low trust (1-3 on a 9-point scale) in the CBC, CTV, Global News and The Globe and Mail, compared to low believers in misinformation (Figure 11). In contrast, high believers were less likely to have low trust in Facebook, Twitter and TikTok, compared to other Canadians.

High believers in misinformation were significantly less likely to report using legacy media (TV, radio, news websites, print newspapers) than other Canadians, and more likely to use Facebook, YouTube and Instagram for news (Figure 12).

Another way of looking at this relationship is the average number of correct identifications of misinformation statements by news source (Figure 13). Those who use legacy media (TV, news websites, radio, newspapers), as well as Reddit and Twitter, for news believe in misinformation at significantly lower levels; while those who use Instagram, YouTube, Facebook, TikTok, Snapchat, Messenger and WhatsApp have significantly higher misinformation belief levels. While most Canadians reported using more than one source for news (83%), and these groups are not mutually exclusive, it is also worth noting that the 17% who said they rely on only one source also had a significantly lower number of correct answers (4.3) compared to 5.2 overall.

Both of these findings echo previous research that found a relationship between consuming news on social media platforms and the propensity to believe in conspiracy theories.¹³ A noteworthy 10% of high believers also said they use no news sources (even with the ability to write-in other sources), compared to just 3% overall.

¹³ Stecula, D., Pickup, M., & van der Linden, C., 2020

High believers in misinformation were also significantly less likely to say they had fact checked something they saw online, and more likely to say they frequently encounter news they believed to be true and later found out was false (Figure 14).

Hate Speech and Harassment

People in Canada reported relatively frequent exposure to online hate speech, though the proportion reporting that they see hate speech online at least a few times per month fell from 48% in 2019 to 41% in 2022. Consistent with past results, self-reported rates of exposure to hate were significantly higher for marginalized communities: 48% among Canadians with disabilities, 50% among racialized Canadians, 58% among LGBTQ2S+ Canadians and 67% among those living in Canada less than 10 years.

About one-third cited examples that may not qualify as hate speech, a finding that provides important context for how survey respondents think about their exposure to online hate speech. For example, 10% cited harmful comments from politicians, 7% cited harmful comments directed at politicians, 7% cited harmful comments regarding COVID-19 restrictions or the convoy, and 5% cited harmful comments directed at celebrities.

When asked if they had ever been targeted with online hate speech that deliberately promoted hatred against a group they identify with, one in ten Canadians said they had; these proportions were significantly higher among marginalized communities (Figure 15).

Overall, 4% reported having their intimate images shared online without their consent. Responses were consistent across genders, though rose to 9% among those aged 16-29.

Another 8% reported having been targeted with online harassment that caused them to fear for their safety. Responses were, again, consistent across genders, but higher among younger people (14% those age 16-29), those with disabilities (14%) and LGBTQ2S+ Canadians (17%). In addition, 6% said they have reported someone to the police for illegal activity online.

One in four (25%) respondents said they have reported or flagged an account to an online platform for sharing illegal content — a proportion consistent with 2019 findings. In addition, 41% said they had blocked or reported an account to an online platform for being fake or automated. Of those who blocked or reported content, 39% rated its effectiveness as high (7-9 on 1-9 scale), 33% rated it as moderate (4-6) and 28% rated it as low (1-3). This relatively positive assessment of effectiveness is quite similar to the findings from 2019, with a slight increase in those rating effectiveness as low, from 23% to 28%. Those who had been personally targeted online with harassment, intimate image abuse or hate speech provided the same overall assessment of effectiveness, with 39% rating the reporting process as highly effective, underlining the importance of these notice-and-action mechanisms for online safety.

Canadians’ Perspectives on Platform Governance

A Role for Government

The findings from our latest survey continue to show that most Canadians are prepared for government intervention to address harmful online activity. We first asked participants to identify who is most responsible for contributing to a rise in harmful online content, and then who should be most responsible for fixing it (Figure 16). While most (48%) still believe platform users are most responsible for contributing to the increase in harmful online content, a clear majority (51%) now believe that platforms should be most responsible for fixing the issue, up 16 percentage points from 2019.

We then asked participants to choose, from among three pairs of statements, which best described their perspective. Findings showed that approximately two out of three Canadians favour platform intervention (Figure 17); these levels are largely stable since 2021. A new statement was added this year to assess views on the role of government in addressing disinformation — again, two out of three indicated that the spread of disinformation is a threat to Canadian democracy and needs to be addressed by government.

Finally, we asked about support for various options to provide recourse to Canadians regarding decisions of large online platforms to remove illegal content, such as hate speech or the promotion of violence. Again, about two out of three (63%) supported a public entity, either a Digital Safety Commissioner that could audit and order platforms to remove illegal content (51%); or an independent ombudsman to help users navigate platform appeals, and investigate and make recommendations regarding platform compliance (38%).(Figures are non-cumulative, as multiple options could be selected.) About half also supported requiring online platforms to have an appeal mechanism for decisions about content or user removal (46%). Only 5% of respondents said no action should be taken, while 11% were not sure.

Support for Action on Online Harms

While Canadians’ views on who is best positioned to lead governance in this space were somewhat mixed, their support with respect to specific policy actions on harmful content was overwhelming (Figure 18). Support levels exceeded 80% for nearly every proposed action. Opposition did not exceed 5%, except with respect to requiring warning labels on false information (7%); applying Canadian laws (8%); and allowing the government to order platforms to take certain actions, such as block or promote certain content or services, during times of crisis with risk of imminent harm, like a terrorist event or public health emergency (15%).

Support for many actions has increased significantly since we last asked the question in March 2021. Strong support for requiring platforms to quickly remove impersonation, violent content, hate speech, repeated spread of false information and bot accounts all increased between 16 and 25 percentage points. It is worth noting that this increase was largely driven by a decrease in the proportion that were neutral to these actions in 2021, as opposition has not changed significantly. Increase in support was also largely driven by those on the left and centre of the political spectrum. Those on the right of the political spectrum had lower overall increases in support, though a majority of those on the right still supported each proposed action.

Views on Private Online Content

A particularly challenging element of designing platform governance is defining which services are in scope for regulatory action. The Government of Canada previously expressed that it intends to exclude from regulation “services that enable persons to engage only in private communications,” and the scope of such an exclusion was a divisive topic among the Expert Advisory Group on Online Safety. Many platforms offer both public and private communication functions, such as large groups, private pages and user-defined ‘close friend’ lists, a situation that adds complexity to the process of defining ‘private communications’.

We have also previously outlined the range of online harms Canadians experience on private messaging platforms in our report Private Messages, Public Harms and advocated that the Government of Canada follow the EU in setting minimum standards for private platforms of a significant size, such as user notice-and-action mechanisms for harmful content and transparency requirements, while not requiring private content scanning or harming encryption. We believe such standards would enable harm reduction, promote greater understanding of online harms, and mitigate the risk of an incentive for companies to create more closed platforms as a means of avoiding new content moderation obligations.

We sought to build on this work by understanding Canadians’ perspectives on the topic. We first asked respondents whether their various social media profiles were set to public or private; we believe this is the first such representative data available in Canada. The majority of respondents said that their social media accounts are set to private, with the lowest proportions reported for Twitter (Figure 19).

We then asked half of the survey respondents which types of online spaces they thought should be required to remove illegal content like hate speech or the promotion of violence. A significant majority of 87% supported content moderation on public pages/profiles, while smaller majorities supported it for private groups (61%) and private pages/profiles (59%). Support fell below a majority (40%) for private messaging groups (Figure 20).

We asked the other half of respondents an open-ended question about what they thought makes an online space ‘private’ within the meaning of the Criminal Code of Canada, which makes it illegal to communicate hate speech “other than in private conversation.” Consistent with the previous question, 36% indicated that they thought no online space should be exempt from illegal content moderation. However, a majority (53%) felt that direct messaging should be considered private, with many describing variations of instances where content’s specific recipients are determined at the time of sending, and differentiating it from private groups where prior content is still accessible to newly-added members. About 20% also mentioned that content accessible beyond a finite number of users should no longer be considered private, with 12% indicating that only conversations between two users should be considered private.