Description

Abstract. Despite its known inadequacies, notice and consent is still the most common privacy practice on social media platforms. Indeed, conceptualizing alternative privacy strategies for the social media context has proven to be difficult. In 2009, Facebook implemented a participatory governance system that enabled users to vote on its privacy policy. However, three years later, Facebook held a final vote that led to the termination of its participatory governance system. Here, we empirically assess this participatory privacy strategy designed to democratize social media policy-making. We describe the different components of Facebook’s participatory governance system, show how users could influence privacy policy decision-making, and report the privacy policies users accepted and rejected by vote. Furthermore, we identify the common themes users discussed during the final electoral period by applying an unsupervised machine learning topic modeling algorithm to thousands of Facebook user comments. Our results demonstrate that users voiced concerns about being insufficiently informed about participation commitments and possibilities, attempted to orchestrate a transfer of the vote to a third-party platform, and engaged in spreading misconstrued data ownership claims. Based on our results, we analyze the key reasons behind Facebook’s failure to implement a successful participation process. Finally, we highlight the significance of framing diversity for privacy decision-making in the context of a participatory privacy strategy on social media.

Keywords: Social Media Democracy, Social Media Governance, Privacy, Online Participation, Topic Modeling. 

Paper