Semaphore: a tool to flag online disinformation
Semaphore is a project created during the 2020 Assembly Fellowship at the Berkman Klein Center at Harvard University. One of three tracks in the Assembly: Disinformation Program, the Assembly Fellowship convenes professionals from across disciplines and sectors to tackle the spread and consumption of disinformation. Each fellow participated as an individual, and not as a representative of their organization. Assembly Fellows conducted their work independently with light advisory guidance from program advisors and staff.
The Semaphore project and this post were authored by John Hess, Michaela Lee, and Isabelle Rice; the project team has backgrounds in software engineering, development, and human rights.
Harmful disinformation narratives, and the communities propagating them, tend to breed in the dark corners of social media sites. While it’s been difficult to make progress on reducing disinformation flows, both the platforms and the research community have worked hard on this problem, including by building workflows and tools to flag inauthentic content. However, the process by which researchers report, or “flag” inauthentic content has a lamentable side-effect: once platforms take down inauthentic content, the researchers who flagged it lose access to the information.
The first step towards a more systematic approach to tackling disinformation is data gathering, and we think an obvious place to start is with the flaggers who are already doing the most practical work. We created Semaphore during our time in the 2020 Assembly Fellowship at the Berkman Klein Center to archive and share data that flaggers already gather and produce.
Semaphore is made of three main pieces of technology: a browser extension, a database, and a website through which the database can be accessed.
In its current beta-version iteration, the Chrome extension, which can be run by anybody, watches for reports the flagger makes on social media sites. (This version of the extension focuses on Twitter.) When Semaphore sees such a report, it grabs the tweet and tweeter IDs and delivers them to the Semaphore database.
The database and the website through which it is accessed are the second and third pieces of technology. The database holds a complete history of the reports that Semaphore flaggers have made. In possible future iterations, the backend would also automatically gather information on whether the reports resulted in the deletion of the reported content as well as archiving that content rather than just storing IDs. The flagger can then visit the site to see a complete history of their reports and the actions the platforms took in response. In future iterations, we also envision enabling sharing of information between research communities, while ensuring that we approach this in a way that encourages productive behavior and avoids potential harassment.
The idea behind Semaphore is to give flaggers the opportunity to do the most good, by giving them the tools to be more strategic and systematic in their fight against the spread of online disinformation. If we continue building the database, this will likely mean sharing data across the Semaphore network and potentially releasing analysis for public consumption. Our top priority is to ensure a safe and productive environment that enables engaged citizens to expose harmful narratives and damaging networks that impact all of us.
For more information on the Semaphore project, visit the team’s website. Learn more about the Assembly: Disinformation program at www.bkmla.org.
