The Future of Automated Journalism: Opportunities and Risks

Abstract

The development of automated journalism has led to a change in the way journalism is produced and labor is divided, but journalists, developers, and editors do not hold uniform views on this development. Among those working in the media there are both concerns and reassurances over the professional implications and quality of journalism. Professionals in the field are encouraged to examine possibilities and consequences from their own perspective given their expertise and involvement. Well-made automated texts are already used to generate news stories from structured data, such as from companies’ stock exchanges. The development of natural language generation algorithms has recently enabled new applications for more complex data, rejuvenating the debate. A report explains one such process, describing how the wording in an article “about earthquakes in Berkeley” triggered the warning “just the dates are the same.” This report broadly sets the stage for this essay to investigate the market penetration of automated journalism, and its implications. The most important concepts for discussion are already mentioned: (1) capabilities, that good automated journalism can only be generated from well-structured and separated data; (2) implications, that automation does not really change anything, human journalists are also not infallible. However, there are further questions left unaddressed, about the extent of automation and what might alter it. The development of this technology can be rather understood and predicted, and a more profound critique, including ethical considerations, is possible.

Keywords automated journalism, natural language generation, media implications, data-driven reporting, professional concerns, ethical considerations, journalism quality, market penetration