Job Title: Robot Reporter
Automation could help media houses survive — and improve journalism
These days it always raises a bit of mistrust, when companies advertise their products using “Corona” in their titles. There is no crisis without crisis winners after all, and not everyone deserves to win. Do editors really need Corona Watch, for example? The newsdesk tool automatically evaluates important sources on the crisis situation and alerts editors and reporters via a Slack Channel. Obviously it doesn’t do so autonomously. Beforehand, the newsroom’s task is to determine sources and selection criteria.
Swedish newspaper Aftonbladet actually regards the tool as a win. Managing Editor Michael Poromaa says, it not only supports his team’s work, but also strengthens their position in the competition with other media: “Before, we were often at most the second to report new cases, now we are the first.” The alternative to using Corona Watch would be for the newsroom to constantly refresh the 21 regional health authority websites, Poromaa says.
Corona Watch was developed by Swedish company United Robots, which offers various automated solutions for newsrooms. Aftonbladet came up with the idea and it was implemented within a day, says Cecilia Campbell. Campbell works for United Robots, previously she co-managed the Reader Revenue group within World Association of News Publishers (WAN-IFRA). Whereas some journalists fear that robot journalism might cost them their jobs, many Scandinavian publishers are ready to embrace it. “News publishers want their journalists to do qualified research-based journalism,” says Campbell, who is Swedish. A variety of simple tasks, however, can be automated.
Many publishers are even convinced that automation can help the industry survive. Stefan Åberg, for example, leads a team of just two dozen journalists at the Swedish media company VK Media. It was impossible to meet all reader expectations with what this team alone could produce, he said at a WAN-IFRA conference in 2019. The solution: “We are building an army of robots.” He reported to have implemented bots for everything: weather, traffic, property sales, sports matches. Since the editorial team had begun using automation extensively, the number of digital subscribers had increased by 70 percent.
Many editorial offices are already using artificial intelligence. So far, it has played the largest role in marketing. Bots automatically offer subscriptions to readers or provide them with personalized content, for example. Some say they listen to the audience better than journalists because they can evaluate data quickly. There are even newsrooms that trust artificial intelligence more than their editors when it comes to news judgement, because at times it seems to be easier to unbias software than to unbias people. Canadian Globe and Mail, for example, use bots when feeding its homepage to ensure that content relating to minority groups is published on a daily basis. In many newsrooms automation is widely used for data journalism projects and fact checking. Others, like British Financial Times or Swedish Dagens Nyheter, use gender bots to check the gender ratio of quoted sources or pictures.
Significantly fewer media companies delegate text writing to robots. The major news agencies were early adopters, since they have to be both fast and broad. This requires lots of routine work. American AP was among the first to have robots report on quarterly company results. Where reporters used to cover a few hundred companies per quarter, robots now manage a few thousand. The Washington Post has massively expanded its election reporting with the help of AI tools.
Will newsrooms face staff cuts? It says a lot about the state of the industry that not even the unions condemn robot journalism as the devil’s stuff. Job cuts and savings programs have been around for years, the advertising crisis following the outbreak of the coronavirus pandemic has sped up the decline of resources. Now it’s about saving what can be saved. The road forward for newsrooms is to do both: serving their customers with exclusive stories researched by reporters on the one hand, and provide them with popular basic reporting on the other. This includes locally adapted weather and traffic news, which no one needs years of journalism training to create. Some of these even generate subscriptions: automatically generated reports on property prices seem to top that particular list.
The sooner newsrooms consider the opportunities of AI and robot journalism, the better. Because editors and newsroom managers who know their way around technology are less likely to get sold on solutions that no one needs. After all, it’s about improving their particular brand of journalism and serving the public better. Flooding readers with ever more stuff doesn’t do the job. Before “hiring” the robots, it is helpful to think about the dos and don’ts. Former AP journalist and advisor Tom Kent has created an ethics checklist for that purpose. Note: AI journalism can only be as good as the data with which it is fed.
Further reading:
- Nicholas Diakopoulos, Automating the News, Harvard University Press, 2019.
- Noam Lemelshtrish Later (ed.), Robot Journalism: Can Human Journalism Survive?, World Scientific, 2018
