Journalists don’t think that automated production of journalism is, in its current form, very useful, a study by Neil Thurman (City, University of London), Konstantin Dörr (University of Zurich), and Jessica Kunert (Ludwig Maximilian University Munich) discovered. The authors introduced 10 British journalists to software capable of automatically producing news content, and observed and interviewed them over the experience.
The participants overwhelmingly saw the technique as being limited, yet having some promise for the future. Automated news production could, when perfected, be used to reduce costs by automating some of the routine reporting and to reduce human error and bias, they suggested.
For now, journalists consider robot journalism to be formulaic, dull, and prone to missing the actual news. As the reporting is based on a pre-determined data stream or streams, context and unexpected turns of events can easily get excluded. The participants also considered the task of setting up the algorithm as tedious and contrary to their professional pride. This attitude may, however, be merely generational, the authors suggests.
The article “When Reporters Get Hands-on with Robo-Writing” was published by the journal Digital Journalism. It is available online from the publisher’s website (abstract free), or from Thurman’s website (open access, PDF file). The article was also covered by Journalism.co.uk.
Picture: Robot by Logan Ingalls, licence CC BY 2.0.