Lou Ferrara, vice president and managing editor at AP, said this new expansion into algorithmic journalism will not lead to a loss of jobs. "We're making sure journalists are being used the best they can," he said.
Ferrara told Journalism.co.uk it would allow AP to report on events they never would have been able to previously, primarily college matches.
"We weren't really covering some of these college events at all in the past but now we're able to do so. We're known for our sports coverage and we want to expand on that."
AP has been using Automated Insights since last summer to cover financial reports, with over 3,000 articles being produced every quarter by the algorithmic program. This has lead AP to hire an automation editor who started earlier this week, their hope being that this will help them expand the use of algorithms within the company.
"We're hoping to move on to other data driven areas," Ferrara said. "Anything data driven like exit polling, the weather, we're looking into using these programs."
Currently stories created by algorithms must have data inputs to be created. A simplified explanation of how it all works is that algorithms, created by Automated Insights, take the data they are given and generate an article following the AP style rules. These can then be published immediately once the data has gone through.
"The reason we're pursuing this right now is because news is moving so fast. [People] are receiving breaking news through Facebook and Twitter and we want to be at the front of that."
But could robots actually run the entire sports desk one day?
Hille van der Kaa, who runs the professorship of media, interaction and narration at Fontys University of Applied Sciences in The Netherlands, has been researching robot journalism.
She explained in an email how systems currently being worked on could one day "take over an entire sports desk".I don’t think a robot will be able to conduct an interview with a human within ten years.Hille van der Kaa, Fontys University of Applied Sciences
"Well, you still need a production team and some geeks, but no traditional journalist who collects and analyses data and produces the story in the end."
One project currently being worked on by Kaa's students is 'Windcatcher'.
"In this project, we put sensors on cyclists to collect information about their well being, for example heart rate. Further, we install a GoPro on every bike to grasp the view of every cyclist. We collect all the data in a second screen application.
"By doing this, viewers can select one cyclist and automatically receive real time information about the cyclist they like on their tablet."
"To go even further, we can make automated reports for different target groups. Using machine learning technology we can teach our algorithm to write an article with viral elements, using specific words or sentence structure, which will go viral on Facebook.
"Or maybe we want a very positive article on a cyclist for a fan site, or a special version for people who are dyslexic."
When it comes to reporting emotion however, Kaa sees difficulties ahead for algorithms. "One of my colleges is working on facial recognition technology which is able to detect an emotion like disgust, measuring facial microexpressions.
"You could use this kind of technologies in automated reporting, but I think there are a lot of cons. Questions arise like: will [athletes] be ever willing to provide this kind of data?
"And do we prefer the analysis of a computer above the spoken words of our favourite sport commentator? I don’t think so. Therefore, we will always have a need for human journalists."
Even though Kaa's research has shown that people equally trust computer writers and journalists, she still doesn't believe that stories which include emotions will be automated in the next 10 years.
"And I don’t think a robot will be able to conduct an interview with a human within ten years, although I once Skyped with a nanorobot and I thought it was so cute that I would probably tell him everything, but that’s another story.
"I am sure we will have a new perspective on the function and meaning of news and journalists within twenty years," she said.
Free daily newsletter
- Newsrooms must learn how to use AI: "Trust in journalism is at stake"
- The 4 types of live video experiments at The Associated Press
- Lessons from AP's experiments with 360-degree video and virtual reality
- How AP wants to help 'democratise virtual reality'
- AP experiments with live streams as appetite for up-to-the-minute video grows