Report Highlights: Processes, People, and Public Accountability

Executive Summary

A report by Chris Tenove and Heidi Tworek examines the complexity of harmful online communication, noting that new factors such as algorithms, interface designs, rapidly changing use patterns, and interactions between digital services complicate the longstanding challenge of determining how messages and media affect people's beliefs and behaviors, making harms difficult to conceptualize and measure even though doing so is crucial for mitigating those harms while balancing interventions against restrictions on free expression.

During the 2019 Canadian federal election, three forms of harmful online communication were observed: abuse of individuals; intolerance and hate toward marginalized groups in public online spaces; and building support for hate in private online spaces, spanning speech that is already illegal in Canada, speech that is harmful but not illegal, and patterns that contribute to systemic discrimination. Identifying online hate speech is difficult for both human evaluators and algorithms because language is dynamic and contextual, and because actors promoting discrimination and hate operate across multiple platforms, meaning single-platform policies are unlikely to be effective.

The study analyzed approximately one million tweets directed at candidates between mid-August and October 31, 2019, and interviewed over 30 candidates or their communications staff. Messages were classified as low negativity (dismissive or disrespectful), medium negativity (insulting or advancing negative stereotypes), or high negativity (hateful language toward social groups, threats, or unsubstantiated accusations of moral or criminal wrongdoing). Initial findings suggest over 40% of tweets fell into one of these three categories: 27% low negativity, 13% medium negativity, and 1% high negativity. Candidates reported that high-negativity messages affected their sense of wellbeing and security and could require time-consuming engagement with police or civil servants, while low- and medium-negativity messages were demoralizing. Campaign teams developed ad hoc, reactive strategies to manage abuse, and candidates called for greater accountability both for individuals posting such messages and for social media companies that allow the abuse and hate on their platforms.

The authors recommend a policy response focused on processes, people, and accountability rather than on individual pieces of content, including creating a social media council to develop principles for addressing harmful communication in the Canadian context, greater support for candidates managing threats, efforts to understand and educate those who promote hate in public and private spaces, and greater access to independent research on platform data.