Key points are not available for this paper at this time.
The broadening reliance on algorithms to generate news automatically, referred to as “automated journalism” or “robot journalism”, has significant practical, sociopolitical, psychological, legal and occupational implications for news organizations, journalists and their audiences. One of its most controversial yet unexplored aspects is the algorithmic authorship. This paper integrates a multidisciplinary theoretical framework of algorithmic creativity, bylines and full disclosure policies, legal views on computer-generated works, and an empirical study of attribution regimes in pioneering organizations that produce journalistic content automatically. Fieldwork included quantitative content analysis of automated stories on 12 websites and interviews with key figures from seven of the organizations that agreed to be interviewed, despite the general reluctance of news organizations to be identified with such an endeavor. The study detects major discrepancies between the perceptions of authorship and crediting policy, the prevailing attribution regimes and the scholarly literature. To mitigate these discrepancies, we offer a consistent and comprehensive crediting policy that sponsors public interest in automated news.
Montal et al. (Fri,) studied this question.