Expanding Platform Boundaries to Counter False Information

Abstract

False information, which can include disinformation, misinformation, and other 'fake news,' is persistent on digital content platforms. Research and strategies for constraining exposure to false information have focused on interrupting its spread within a platform, rather than its initial entry into a platform. We propose expanding the boundaries of the platform's functions to include monitoring and identifying false information internet domains. There are few burdens on establishing a domain, which has the unintended consequence of making it quick and inexpensive to create, replicate, and distribute false information through domains. To counter some of these advantages, we examine whether a platform can identify problematic domains before users engage with the domains' content on the platform. To establish the feasibility of such a strategy, we show that (1) a platform can predict false information domains at the time of registration, (2) the predictions are valuable under different platform governance types, and (3) the platform can exploit a false information creator's low transparency to sustain the effcacy of the prediction models.