Contentpepper® Blog

Three Cases for Artificial Intelligence Usage in Digital Publishing – what we can learn from big media houses

2019 February 25 at 13:17PM

Artificial Intelligence is becoming an increasingly important buzzword when it comes to the future of digitisation. Intelligent systems are on the increase in the digital publishing business. In the Reuters Institute's Digital News Report, 72% of editors surveyed were willing to experiment with Artificial Intelligence in their newsrooms.

The fields of application are manifold: from the automation of the editorial processes through research to the interaction with the readers - Artificial Intelligence already offers a multitude of options that could permanently change digital publishing. But how exactly can intelligent systems fit into editorial everyday work and what concrete benefits do they offer?

 

#1: Automated Journalism

The Washington Post published over 850 articles last years, generated not by human endeavour, but by algorithms. Meet “Heliograf”, the intelligent, in-house system, which now acts as a robo journalist. At the Olympic Summer Games in Rio, the system was already in use to evaluate the results of various disciplines and pack them into short, understandable texts. Meanwhile, the areas of application have been extended, for example, to smaller football games or governor elections. While the robot takes care of the analysis of data and the writing of small reports, journalists can focus their energy on larger background topics. Publishers like Yahoo Sports and American news agency Associated Press are also taking advantage of automated journalism.

However, the abilities of Automated Journalism are not exhausted by short message writing. More and more media houses, such as USA Today, Times Inc. or Reuters, are cooperating with the Wibbitz platform. Wibbitz is an application that automatically summarises texts in short videos. This gives publishers the ability to automatically produce and play news in an additional format. Users can then consume the content as text or video at will. Automated Journalism not only supports media houses in creating simple, routine content, but can also contribute to the variety of journalistic offerings.

#2: Reader Interaction

Chat bots have already established themselves as a means of customer service in many industries. Now in digital publishing, artificial intelligence offers new opportunities to interact with readers and to respond to their needs. The Guardian has developed an interactive reader tool that allows users to get the latest news via chat on Facebook Messenger. Users control settings such as when the chat bot should deliver the news to them, as well as which region they are interested in and which topics are relevant. The bot also responds to messages from readers with other relevant content. In a nutshell, users no longer have to scroll through news sites; instead they receive messages tailored to them via a personal channel, with the option of further information by requesting directly through a message.

The New York Times, on the other hand, is experimenting with a tool that could radically change the moderation and support of online commentary. “Perspective API”, an application from the Google company Alphabet, automatically sorts comments based on their pitch and places them on a scale. Thus, moderate and neutral contributions are separated from those that of an aggressive or even insulting nature. Users can decide for themselves which kind of comments they want to see and which they prefer to be hidden. Not only could this feature improve the user experience, but it would also take a workload off the New York Times comment moderators by automatically filtering and sorting out insulting and offensive comments.

#3: Semantic Discovery

Another field of application for Artificial Intelligence in digital publishing is that of so-called Semantic Discovery, whose game forms have already prevailed in various media houses around the world. Its goal: to automatically correlate, tag and evaluate data, news and other relevant content to enable a more efficient use of news sources.

Using the AI tool "Juicer", the BBC has implemented one of the most comprehensive tools in the field of semantic discovery. The application collects information from around 850 news outlets and feeds worldwide, tags them, and places them in one of the categories People, Organizations, Places, and Things. In the search, journalists can access this comprehensive database, keep track of the overview of related and similar articles by using classification, and resort to comprehensive, artificially generated information pools. In future, "Juicer" could also help readers: In articles certain words could be linked, so that when clicking on them, windows with background information pop up.

Conclusion

The importance of Artificial Intelligence for digital publishing is growing rapidly. Leading media houses around the world are beginning to discover and use the power of AI tools. Be it for the automation of routine tasks, for the support of readers or as a research facilitation for journalists and other copywriters - artificial intelligence could decisively shape the future of the media industry.

Want to see more? Find insights from the publishing industry in our free White paper „Publishing in the Change – Digitisation, Crossmedia and Technology in the publishing sector“.