Reference:
Parratt-Fernández, Sonia, Javier Mayoral-Sánchez, and Montse Mera-Fernández. 2021. “The Application of Artificial Intelligence to Journalism: An Analysis of Academic Production.” Profesional de La Información 30 (3). https://doi.org/10.3145/epi.2021.may.17.
This paper presents a summary of academic research on AI use in journalism, based on the authors’ review of 358 texts published between 2010 and January 2021. The materials they reviewed were found through academic databases including Scopus and Web of Science, in addition to Google Scholar. Most of the articles were published in English, and the majority was from the United States.
Given significant developments with AI, and AI in journalism, since 2021, this paper is really a snapshot of research published for the period covered. The authors do note a rapid increase in research until 2019, with a dropoff in 2020 presumably from disruptions of the COVID-19 pandemic.
Most articles the authors reviewed were conducted using qualitative research methods, and focused on data journalism, writing by robots, and verification of news. The find gaps in other topics, including how AI affects the role of journalists, content personalization, and coverage of AI in journalism education.
The authors identify the year 1952 as the first instance of computer use in news production. The terms used within the journalism profession have changed as new technologies were developed:
- Computer-assisted reporting (CAR) from the late 1960s to early 2000s.
- Data journalism became a common term in the late 2000s.
- Computational journalism in the mid-2000s.
- Automated journalism, or algorithmic journalism, or robot journalism are currently-used terms today.
I think the language is important but the terms often overlap. Based on my experience and reading of current journalism texts, the terms CAR and data journalism continue to be used. That said, the term “automated journalism” is rising quickly as generative AI applications gain ground in newsrooms.
From the materials they selected, the authors identify the most prolific researchers in the field of AI journalism (I just made that my favorite term):
- Nicholas Diakopoulos (United States)
- Neil Thurman (Germany)
- Seth C. Lewis (United States)
- Ester Appelgren (Sweden)
- Eddy Borges-Rey (Qatar)
- Meredith Broussard (United States)
As noted above, the majority of the research reviewed used qualitative methods, often based on case studies and interviews. Few used discourse analysis (only 3) or focus groups. The authors suggest that quantitative research may be relatively neglected.
The majority of articles were focused on how AI is used in newsrooms (e.g. data journalism, automated writing, verification), and few were focused on ethics (“The ethical issues arising from the arrival of artificial intelligence in the media have not yet attracted much interest”, p.8) or effects on employment of journalists.
Importantly, the preponderance of articles showed a positive view of AI in journalism, and argue for AI adoption and adaptation. The authors state that “a positive view of the effects of the advent of artificial intelligence in the media has predominated in the research papers reviewed to date. In fact, almost no articles exist which voice a negative view towards these new technologies” (p.8). Although they do note a trend toward more neutral stances among researchers.
A major point of the paper is to identify research gaps, and they point to several:
- Rethinking the journalist’s role
- Personalizing content for specific audiences
- Training needs
- The reformulation of work in newsrooms
- Effects of AI on journalism jobs
Adventures in Information Science 