Disclose, decode, and demystify: An empirical guide to algorithmic transparency

The values, practices, and standards for algorithmic transparency are still developing, especially regarding questions of how and when information concerning algorithms can and should be disclosed. The focus group discussion analysis of the three case studies provided us with an empirically grounded typology covering a broad spectrum of information disclosure for various algorithms. As moderators guided and prompted the discussion about algorithmic transparency, participants collectively gathered, categorized, and displayed possible avenues for information disclosure. In the research literature, journalistic transparency is frequently discussed as an ethical imperative and as a way to increase legitimacy and credibility in the news media. The last stage of the input-output pipeline - the interface - directly interacts with end users to provide transparency information in a comprehensive form. As transparency is increasingly becoming a core concept for journalistic accountability and journalistic truth telling, the discovery and recommendation of feasible transparency practices is critical for improving journalism’s credibility and legitimacy.