.svg)
Municipal elections in Canada are the bedrock of our democratic system - vibrant local governments contribute to a healthy political culture by encouraging civic engagement, fostering a sense of community and nurturing the values and practices of democratic governance. Understanding the role that digital abuse is playing in our local democracies is critical for understanding the health of Canadian democracy writ large. To shed light on this, the Samara Centre for Democracy turned to our SAMbot project, a multi-year machine learning initiative that measures abusive content received by Canadian political candidates online.
Using SAMbot, we tracked online abuse in eight municipal elections held in 2022 in the following cities: Vancouver, Surrey, Ottawa, Brampton, Toronto, Winnipeg, Yellowknife and Charlottetown. Across these elections, SAMbot evaluated over 465,000 tweets received by 524 Twitter accounts belonging to city council, school board, and park board candidates along with political party accounts.
Across eight elections we found more than 86,000 abusive tweets with high volumes of online abuse in the Ottawa, Toronto, Vancouver, Winnipeg, Brampton and Surrey elections. Our previous SAMbot reports established that online abuse is a significant problem in federal and provincial elections. These new findings reveal that digital toxicity is affecting our municipal elections as well.
As expected, abusive tweets were common in mayoral races; however, school board trustee candidates also received some of the highest volumes of abusive tweets, many in the form of threats and identity attacks. Identity attacks were, in fact, a significant form of abuse in a number of elections, with attacks concentrated to specific candidates. High levels of identity attacks particularly affected local online conversations in Ottawa and Brampton.
In the Charlottetown and Yellowknife elections we found limited candidate presence on Twitter (and therefore limited abusive content), which suggests that the online political conversation in those locations is taking place on other platforms. This points to the need for more robust research into abuse across different social media platforms.
The volumes of abuse detected with SAMbot illuminate the challenging working conditions experienced by municipal candidates on the campaign trail. These working conditions, facilitated by digital technologies, threaten to reduce participation and representation in our democracies. Their effects are crucial to understand as we work to build a robust, responsive and inclusive democratic culture in Canada.
Throughout the fall of 2022, the Samara Centre for Democracy deployed SAMbot in eight municipal elections across Canada to measure abusive content received by candidates on Twitter. Over 6.5 million people live in these eight municipalities, which is approximately 17% of the Canadian population.
As of 2022, Twitter is the fifth most popular social media platform in Canada with 40% of online Canadian adults possessing a Twitter account. Among these users, demographics tend to skew younger, with higher household incomes and educational outcomes.
.png)
The SAMbot project uses AI for civic inquiry in order to better understand technology’s influence on our democratic culture. As political discourse is generally at its most toxic during campaigns, SAMbot helps us gain critical insight into the current state of online Canadian political conversations. SAMbot has previously tracked federal and provincial elections. These municipal deployments allow us to measure the types and extent of online abuse in local elections.
Across these elections, SAMbot evaluated over 465,000 tweets received by 524 Twitter accounts belonging to city council, school board, and park board candidates along with political party accounts. More than 86,000 abusive tweets were found.
SAMbot uses a machine learning model to assess abusive language. These models are ever-evolving which means that each time SAMbot is deployed, it becomes more accurate and informed. In addition to measuring volume of abuse, SAMbot also provides insight into the type of abuse received by candidates (e.g. identity attacks, sexually explicit content, threats).
This report presents methodological details and election-specific analysis.
SAMbot tracked each of these elections, in their respective regions & time zones, from the evening that the nomination period closed to the evening of election day. This information was drawn from each municipality’s website and/or provincial election offices. Candidates were only tracked by SAMbot if they had a public and active Twitter account by the end of their election’s nomination period.
SAMbot is a machine learning bot. Each message that SAMbot tracks, whether that’s a reply, quote tweet or mention, was analyzed and scored on five abuse categories:
Tweets can fall into two or more of these categories simultaneously. For example, a tweet can be both a threat and an identity attack.
If a tweet is scored as any of those five abuse categories, it would also be counted as an abusive tweet. Whether a tweet meets just one, or multiple abuse categories, it is counted as a single abusive tweet.
Click here to learn more about SAMbot’s methodology.
We thank the team at Areto Labs for their technical expertise, input and collaboration.
How to cite: The Samara Centre for Democracy, Online Abuse in Local Elections: The SAMbot Municipal Report, (Toronto: The Samara Centre for Democracy, 2023), www.samaracentre.ca/project/online-abuse-in-local-elections-sambot-municipal-report.
.png)
-min.png)
.jpg)
.png)
-min%20(1).jpg)
.jpg)
.jpg)
.jpg)
.jpg)
