© 2024 WNIJ and WNIU
Northern Public Radio
801 N 1st St.
DeKalb, IL 60115
Northern Public Radio
Play Live Radio
Next Up:
0:00 0:00
Available On Air Stations

How does WNIJ use artificial intelligence in the newsroom?

Image by Engin Akyurt from Pixabay

No, robots do not write our stories at WNIJ. We don't use fake bylines either. But we do use some applications in our newsroom that help our reporters work more efficiently while making sure our stories are discoverable on the web.

Increasingly, you may be reading about instances of news organizations using artificial intelligence that have yielded unintended results that erode trust in how journalists produce news. 

CNN: Gannett to pause AI experiment after botched high school sports articles 

We want to be transparent about our use of tools that involve artificial intelligence and how we make the decision to use them.

In May, the Radio Television Digital News Association issued guidelines for the use of Artificial Intelligence in journalism.

Among them, a section on transparency with the following considerations:

  • Does the benefit to the public of your use of AI outweigh any risk or detriment to trust in news by not disclosing its use?
  • How does your audience feel about the use of this technology?
  • What falls under the definition of AI? Examples: content generation vs. content distribution/organization vs. video or text editing vs. grammar/spelling tools.
  • Where will you provide disclosure about your use of AI? Examples: Website, social media, on-air
  • Are journalists able to review any and all AI-influenced content before it reaches the consumer? If not, can you justify its use to the audience?

What do we use at WNIJ?


Reporters used to spend hours for a process called "logging tape." It involved manually transcribing every word a person said so that we could find where the soundbite was that we wanted to use in our stories. For the past few years, WNIJ reporters upload raw audio into Otter.ai which provides an automated transcription of minutes or even hours of audio interviews within a few minutes. It isn't a perfect transcription. We like to joke internally about “Otter-isms.” These are botched names or phrases that it generates when it can’t entirely understand what someone said. While it gives us instant access to finding the soundbites we want to use, humans need to make sure to check spellings and log the direct quote we plan to use word for word to weed out those “Otter-isms” before publication.


WNIJ also uses a Slack integration called YESEO that was developed by Ryan Restivo during his fellowship at the Reynolds Journalism Institute. It reads text either before or after publication to suggest keywords and headline suggestions. A pilot version allows for social media text suggestions to increase the chances that people will find our stories on social media. We do not automatically choose a headline just because YESEO suggests it, but it does give us some options to consider, or a head start on keywords to include. We have seen results in reach and views when we include suggested keywords as we are building the web posts. Still, humans in our newsroom make the final call on how exactly we want to present information that we have gathered.


Cameras now allow you to “magically” erase a person with the slightest swipe. You can also summon a graphic with just a few keywords using AI.

If we use AI generated images to accompany any of our content, you will see that clearly labeled in the caption.

Regarding news photos, it would be unethical to remove a person from a photo that we publish to accompany news stories. It is our role to bear witness to history and we shouldn't alter history.

The audio editing software we use called Adobe Audition has long allowed us to enhance audio that may have extraneous hiss, pops/clicks, or in need of overall restoration. We recently tested an AI audio enhancer to demonstrate to students just how easy it is to manipulate audio and why “just because you can, doesn’t mean you should.” For the demonstration, our reporter had a speech they had recorded from pretty far away. The AI tool tried to “fix” it to the point that it artificially generated garbled and nonsensical words for what it couldn’t understand from afar. It was the audio equivalent of AI’s six-fingered hand problem.

Final thoughts

While AI is a very powerful tool, without human intervention or ethics in place, it will fill in gaps on its own — and that is where we draw the line.

You can trust that when you see a story populate on wnij.org that the newsgathering, editing, writing, and audio production has been done by a real person in this local newsroom.

Jenna Dooley has spent her professional career in public radio. She is a graduate of Northern Illinois University and the Public Affairs Reporting Program at the University of Illinois - Springfield. She returned to Northern Public Radio in DeKalb after several years hosting Morning Edition at WUIS-FM in Springfield. She is a former "Newsfinder of the Year" from the Illinois Associated Press and recipient of NIU's Donald R. Grubb Journalism Alumni Award. She is an active member of the Illinois News Broadcasters Association and an adjunct instructor at NIU.