The One Thing Journalistic AI Just Might Do for Democracy

Abstract

A recent wave of enthusiasm surrounding artificial intelligence (AI) has seen many observers focus on its potential to transform journalism, as with other forms of knowledge work. This conceptual essay, however, attempts to steer the conversation away from a technocentric emphasis on applications and instead pivot to the fundamental but overlooked issue of normativity: in this case, what should AI actually do to help fulfill journalism’s broader democratic aims? Building our analysis on Rasmus Kleis Nielsen’s approach that emphasizes one essential thing that journalism just might do for democracy—namely, provide accurate, accessible, diverse, relevant, and timely news about public affairs—we argue that journalistic AI could be normatively imagined and evaluated with a similar focus on simplicity and dose of democratic realism. In connection with three primary stages of newswork—information gathering, selection and production, and distribution and consumption—we explain how journalistic AI not only can work but also should work in the service of accuracy, accessibility, diversity, relevance, and timeliness. Ultimately, by placing normativity at the center of this analysis, our goal is to explore how AI might be oriented to enhance how news performs in each of these domains and thus make a pragmatic difference for journalism and democracy.

Keywords:

• Accessibility
• accuracy
• artificial intelligence
• diversity
• journalistic AI
• normativity
• relevance
• timeliness

Acknowledgements

We thank Matt Carlson, Laurence Dierickx, Kai-Ti Kao, Seungahn Nah, Thuy Nguyen, Rasmus Kleis Nielsen, Sue Robinson, Eleanor Sandry, Keya Saxena, Sonia Jawaid Shaikh, Felix Simon, the journal editors—including Oscar Westlund and his editorial team as well as the guest editors of this special issue—and the anonymous reviewers for providing helpful feedback on earlier versions of this article.

Disclosure Statement

No potential conflict of interest was reported by the author.

Notes

1 Even as this paper takes up each of these elements in Nielsen’s (2017) formulation, we are setting aside, for the sake of brevity and clarity, a particular focus on the norm of news as “independently produced.” This qualifier about journalistic autonomy is no doubt important, but it falls outside the scope of our analysis because it is not so clearly connected to the role and function of artificial intelligence in journalism (for a discussion of autonomy and journalism, see Sjøvaag 2013; for an examination of how news organizations might be potentially subject to platform companies because of the increasing use of AI, see Simon’s [2022] article for this special issue). Future analyses, however, may want to take up more structurally oriented issues of independence in connection with technology generally and AI particularly.

2 There is at least one important exception to this: Given how the development and deployment of AI in journalism raises questions about the role and influence of technical specialists, it should be noted that such individuals in news organizations are predominantly white and male (e.g. see Usher 2019). Achieving greater diversity in content presupposes achieving greater diversity in the media workforce or in media ownership.