Machines and humans don’t always seem like natural allies, but that is changing as newsrooms learn how robots and journalists can work together.
During the third day of the INMA Media Subscriptions Summit on Tuesday, Cecilia Campbell, chief marketing officer for United Robots, explained in a workshop how emerging technology fits in the context of newsrooms and how it’s changing processes.
“Newsroom automation is, of course, a much bigger topic than just automating the content production. Artificial Intelligence is used to automate a number of processes in newsrooms today,” she explained. “Some are internal, like news alerts and rules to help with things like tagging others directly, and others affect the reader experience, such as personalisation or content recommendations.”
When robots write text, it happens in one of two ways, Campbell explained: NLG, or natural language generation, which uses AI to generate content using structured data, or machine learning-based language generation, which uses AI to create human-like text.
While these approaches have plenty of advantages, Campbell said it’s important to understand the risks as well: “The biggest risk we see is inconsistent irregular or incomplete data. So the content is only as good as its source.” Newsrooms must also consider how and where they publish automated content she said. “If it’s done badly, it could damage your brand.”
Newsrooms employing such automation are discovering the sweet spot of how reporters and robots can work together, Campbell said. Robots can write news stories based on data coming in, which allows the newspaper to break the story. Then, reporters can follow up and provide more depth and information. So when approached properly, there’s a place for both in the newsroom.
NTM, a local Swedish media company with 18 newsrooms in the country, began working with United Robots in 2016. Jens Pettersson, chief digital reader revenue officer for NTM, explained how NTM is using robots to improve coverage. Robots, he said, are extremely fast and accurate, so they are ideal for covering topics such as real estate, sports, traffic, and new businesses. Next, NTM will use robots to provide weather reports.
“We use this primarily to speed up production and create a lot of high volume on the right topics and let human journalists focus on other stuff that is more important for them to focus on,” he said.
Anna Karin Tilleby, project manager/content development at NTM, said the robots have taught them several important lessons: “In the beginning, reporters feared they would lose their jobs to robots,” she said. That has not been the case. Now, they see the robot-generated content as an add-on to reporting — and journalists can focus on doing what they do best.
“You need to analyse and adjust,” Tilleby said. “Invest time in optimising the use of content [such as] when and where to publish, pageviews, spent time and conversions.”
Finally, she said, it’s important to simplify the process: “Almost everything is published automatically on our sites. One of the big benefits of using this material is that it's time-saving and efficient.”
And, Pettersson added, it’s effective: Robots are responsible for 952 new subscriptions, which accounts for 2.5% of NTM’s total subscribers. And robot-written copy has received 9.4 million pageviews from subscribers, or 4% of total pageviews.
Cynthia DuBose, managing editor for audience engagement at McClatchy, explained how the company approached the use of automation with local content.
After realising that many of the topics driving new audiences to them revolved around information, DuBose said it was apparent they didn’t have the manpower to cover many of the topics — nor did they want to dedicated reporters to do so.
“We want our journalists focused on the journalism, not tracking down like many how many homes were sold in this zip code every week or tracking down a lot of sport scores. We wanted them focused on what they were skilled to do, and that's the journalism.”
That led to using AI projects in seven of the 29 markets it is in, and that number probably will grow to 10 by the end of February.
One critical component publishers shouldn’t overlook is having a champion for the effort. That champion will help address concerns about AI because DuBose said people will naturally be nervous about the idea.
“They need to be curious; again, we’re journalists. So maybe skeptical, too,” she said. “They need to be able to ask questions and be comfortable saying, ‘Tell me again, how is this going to work?’ Because that’s the person who, when other people in the newsroom say, ‘What are we doing? Why are we doing this?’ [they] are able to answer those questions.”
The Summit continues Tuesdays and Thursdays through February 15. You can register here.