For Tech Companies

Close anonymity loopholes . Facebook Groups and Pages can be created without linking to a specific Facebook user; only an email address is necessary. Admins can post to their groups under the Group or Page name, effectively allowing anonymous, or at least, pseudonymous, posts. Ending anonymity will not end harassment; in some cases, harassers posted on Facebook under their legal names anyway. But many targets suspected that local officials and community leaders were behind these anonymous groups, who were able to avoid public censure or responsibility for their actions. If Facebook requires users to use their “real names,” why are Pages and Groups not subject to this rule? This loophole is doubly frustrating for the users on these pages who cannot be anonymous as they endure harassment—policies should clarify situations in which anonymity is no longer permissible, like when harassment occurs. The bar should also be higher for public figures than for private citizens; civic leaders should not be able to hide behind anonymity on community social media groups.

Use network analysis to detect local harassment campaigns . Facebook uses network analysis to identify campaigns in other contexts, but does not appear, from what we found, to apply the same methods to local Pages and Groups. All the targets we spoke with said they reported harassment to Facebook multiple times, but were told the individual posts didn’t violate Facebook’s Community Guidelines —a clear downside of our digital ‘public spaces’ being controlled by private companies. Most platforms review reported posts and comments individually, even though the context is often necessary to understand the harassment, such as multiple harassing comments that accumulate, or implicitly hateful language and references. Facebook and other platforms should also use multiple reports in the same group or against the same user as a signal that human review by subject matter experts is needed.

Provide user-friendly tools to report & prevent harassment . Our Abuse Reporting Scorecard lays out specific recommendations for fixing abuse reporting tools, while our Social Pattern Library demonstrates how these tools would work. For local online harassment campaigns, we recommend batch reporting, real-time support for targets, related activity/cross platform reporting, and delegated access, among other key tools. Targets also need better tools for documenting abuse to ensure effective reporting. Most interviewees took screenshots, but these did not always include necessary information such as user IDs, dates, or comment thread histories.

Make transparency reports accessible and understandable to non-experts. Users should be able to make informed decisions about the risks of social media, including how well the platform enforces its rules against hate and harassment —especially to contextualize their own experiences with localized online harassment. Although the majority of users may not read these reports, they should be aware of their existence and able to make sense of the insights. Reports or hubs should be easily discoverable (reachable in one click from the landing page or from a relevant policy page) and include clear explanations of key metrics and their significance. Making transparency reports more prominent also sends a strong signal about a platform company’s commitment to trust and safety.

Give greater access to data for researchers and third-party auditors. Academic and civil society researchers can be partners to tech companies in identifying gaps in policy enforcement and making online social spaces safer, and in deepening understanding of localized online harassment. Independent auditing is valuable for building trust with the public and holding platforms accountable for addressing harms. That work is only possible when researchers have access to platforms’ data—typically via open APIs—in a responsible way that protects user privacy. Yet recently most social media platforms have been moving in the opposite direction by restricting access to APIs.

ADL Center for Technology and Society • Hate Thy Neighbor: Online Hate in Local Communities