The next speaker in this ECREA 2014 panel is Chris Anderson, who directs our focus to the journalistic audience as a raw material for algorithms. Historically, audiences were first constructed as professionalised: they were insulated from journalistic practice. Later, a dialogue understanding of the audience saw it filtered through national issue forums (e.g. town halls), and became a discursive participant; finally, the audience was seen as an active in both politics and the media, especially with the arrival of Internet-based communication technologies.
We may now see the emergence of a concept of the algorithmic audience: this emerged with the Demand Media model, which created online content based on a combination of predicted audience demand and expected return on investment – here, algorithms identified potential topics, calculated the potential total value, determined compensations for authors, and farmed out the writing jobs to an army of freelancers (for a total of some 7,000 articles per day).
But this model was suitable for the search engine era, not for the social media-centric post-search era. Today, algorithms seek to select the news items that are seen as most relevant to their interests; this creates a kind of cybernetic audience by focussing on pre-known needs of users, generates a balanced feed that addresses these needs, and reacts responsively to changes in these audiences. What does this mean for the functioning of democracy?
