Media Accountability in the Algorithmic Age - Portugal and The Netherlands: Two Media Systems in Comparative Perspective

The widespread use of algorithms in several areas of society raises concerns about their impact on the lives of people (World Wide Web Foundation, 2017). Particularly in journalism algorithms are used in several ways. For instance, to distribute content on social media, to find topics to write about, and to rely on data-analytic tools (Diakopoulos & Koliska, 2016). Owing to this, some scholars have been urging for algorithmic accountability in journalism (Berners-Lee, 2018; Diakopoulos, 2015, 2017; Diakopoulos & Koliska, 2016; World Wide Web Foundation, 2017). Within the two main media systems in Europe this thesis asks how can journalists in Portugal and in The Netherlands be held accountable for publishing their content on algorithmically-controlled environments and using algorithms in their work? Based on the different expectations concerning the role of journalism in the Polarized Pluralist and Democratic Corporatist models (Hallin & Mancini, 2004) this research assesses potential solutions for algorithmic accountability within two opposing frames of accountability – the political and the professional frame. Through expert interviews, this inquiry shows different possible solutions for Portugal and The Netherlands concerning algorithmic accountability in journalism. In relation to Portugal, the main findings show that governmental intervention should be avoided, except for the case of the use of social media. One of the solutions proposed is to monitor the content that is published on these platforms by news media companies for normative purposes. Although self-regulation is considered the most adequate path, the structure of the news media market might hinder algorithmic accountability as well as some features among the class. One of Surprisingly, one of the solutions points to the market frame, not in the scope of the research. Differently, the experts from The Netherlands give a more nuanced approach on the matter. Although self-regulation is deemed ideal, a more indirect governmental interference is considered positive to hold journalists accountable, mainly through governmental pressure and State subsidies to independent monitoring organisations. Indeed, the inputs lean to the public responsibility frame rather than the political frame, which does not comprise the focus of the research. In both countries, algorithmic transparency is considered a possible path to hold journalists accountable, but it needs 3 to be thoroughly assessed and implemented. The findings also indicate that algorithmic accountability is already happening in both countries, mainly through informal self-regulation mechanisms. On the whole, this thesis aims to spark the discussion not only in Portugal and in The Netherlands but also in other Polarized Pluralist and Democratic Corporatist countries concerning algorithmic accountability in journalism. Most of all, it informs further research, media policy makers, news media companies, and journalists themselves on possible paths to algorithmic accountability.