Automated journalism: AI already writes news, verifies it and offers personalized information


The irruption of artificial intelligence has given rise to Automated Journalism , that is, the automatic writing of news with minimal or no human input using natural language computing techniques.

We are facing an emerging reality that occupies all stages of the journalistic chain, mainly three phases: automation in the collection of information, automated content production (news writing) and information distribution, and relationship with the audience.

The automation of journalism is already a fact

Over the past decade, the application of AI in journalism has spanned from news agencies to mass media on a global scale. For example, in the case of US media, about a third of what Bloomberg News publishes uses some form of automation. Forbes has a tool called Bertie to help journalists improve their copy, and The Associated Press news agency uses Automated Insights’ Wordsmith platform for automated news writing.

In European media, the BBC uses an online platform that allows journalists to design their own automated news templates. Le Monde used automatic copywriting to cover the 2015 elections, and a bot called Tobi produced almost 40,000 stories about the 2018 Swiss elections.

In China, the Xinhua agency developed the first artificial newscaster

With the automation boom , companies specializing in the creation of natural language generation systems for writing journalistic texts are flourishing, most of them through labs or start-ups .

In Spanish media

In Spain projects have been appearing progressively. Medusa is a pioneering example, from Vocento, that uses automated journalism to generate information on the situation of beaches and ski slopes.

Also noteworthy are the AnaFut bot from El Confidencial, for the automatic writing of sports reports; audiovisual automation at RTVE , which, together with the EFE news agency and the Narrativa start-up , is carrying out a pilot project for the automatic writing of texts; or the Gabriele software created by Narrativa to write journalistic pieces automatically, and which collaborates with various Spanish media.

Although the implementation of artificial intelligence in newsrooms is still in its infancy, it has begun to arouse significant business interest as a new business model.

The technology is ready, but trust is lacking

The media feel mistrust and ignorance about the very application of AI to their processes.

This explains its slow implementation, as we have verified in a study by the University of Valladolid , carried out by a hybrid team of Journalism and Telecommunications researchers. The study includes the first catalog of AI tools and services for the communication sector.

We have found that there is “sufficient technological muscle made in Spain ” for software development applicable to the entire journalistic process, in all formats and platforms. In other words, the entire process can be carried out , from start to finish , in a fully automated way.

In the case of the information gathering phase, AI can analyze large amounts of data and find patterns , facilitating the search and verification of sources and audience analysis.

Newtral already uses AI tools for news verification and Paradigma for audience data analysis.

The automated news production phase is the one in which the fewest tools are found, and the least applied by Spanish companies.

Some examples are the case of the software Gabriele de Narrativa or Leo Robot from Dail Software for the automatic writing of news through algorithms.

How exactly does the news we are looking for reach us?

The use of AI for news distribution is one of the most widespread. It allows personalizing the contents and linking them with the audience, which allows them to retain their loyalty, and the monetization of information.

The media are concerned above all with reaching segmented and potentially interested audiences. Among the tools observed in the study is PiperLab , which uses AI for audience segmentation analysis; Tools for content curation and Sngular . for the analysis of social networks or trends, among others.

AI = efficiency for media in crisis

Despite the concern of many journalists that the application of AI could put their jobs at risk and the reluctance of many other media outlets to introduce these tools in their newsrooms, the main results of the study show that it is a development that can be a key boost in the sustainability of the media. Both for the improvement of the efficiency of the processes and for the effective loyalty of the audiences.

According to the innovation managers consulted, the application of AI brings economic benefits, as has already been verified in other business sectors, as it is mainly focused on optimizing journalistic functions .

In addition, AI helps reduce the workload by automating repetitive and tedious tasks such as news transcription , fact-checking or comment filtering .

AI also makes it possible to lower costs and increase audience loyalty .

The experts consulted draw an “irreversible reality” in which technology can help produce higher quality journalism, and also a “forced change” that goes beyond the journalistic sector itself.

The transversal technological and social transformation not only requires great headlines, but also a determined implementation.

Author Bios: Pilar Sanchez-Garcia is a Professor of Journalism, Alba Ten-Grace is a Predoctoral Researcher, Carla Calvo-Barber is also a Predoctoral Researcher and Noemi Merayo Alvarez is a University Professor all at the University of Valladolid