When Privacy Goes Viral: Rethinking Freedom of Speech in Ageof YouTube and Instagram

Social networking sites, such as Facebook, are modern-day agoras, where public debate takes place. Social media freedom of speech has thus become a problem, and calls for better regulation have been made. Public discourse is all about content moderation, as needed to remove harmful content by some, but censorship by others. What we have here is that we argue the modern debate is centrally focused on the speaking aspect of speech but overlooks a powerful way in which platforms have started invading free speech on the side of the audience. Rather than simply speaking to one's network of adherents, speech on social media is now organized by algorithms with the intent of maximizing user engagement and commercial appeal in the interest of targeted advertising. The result is that speech audiences are algorithmically determined, and it is a trend we term 'algorithmic audiencing'. We introduce algorithmic audiencing as a discovery, a new trend gone unnoticed so far. We show that it impacts free speech in fresh and unforeseen ways not possible in pre-digital times, by amplifying or suppressing speech for profit, thus distorting free and equal exchange of ideas in public discourse. If black-boxed algorithms control who we talk to, the hurt parties, free speech shifts from 'what can be said' to 'what will be heard' and 'by whom'. We must problematize the audience side of speech urgently if we are to fully understand, and master, free speech on social media. In the context of Information Systems research, algorithmic audiencing opens up entirely new research dimensions.