At this year’s LMA Fest, the AI Community Journalism Lab showcased real-world experiments proving that artificial intelligence (AI) has the potential to create efficiencies in the newsroom.
The AI Lab, made possible with funding from Walton Family Foundation, has helped 21 publishers explore the possibilities of AI to free up more time to cover local news, build their digital capabilities, and serve their communities. For example, experiments from The Durango Herald, Southeast Missourian, The Baltimore Times, and Shaw Media are demonstrating how AI-powered tools are helping small teams in a big way, including, as one publisher found out, responding to local breaking news.
Designing and conducting experiments and creating community outreach opportunities are the main activities of the AI Lab. The experiments provide the publishers with a structured way to carefully test AI while also advancing their experiences with emerging technology.
(From L-R) Session moderator John M. Humenik with Jon Rust, Rust
Communications; Paris Brown, The Baltimore Times; and John Sahly,
Shaw Media
John M. Humenik, program director of the AI Lab, led a discussion at LMA Fest with three of the publishers who are conducting AI Lab experiments, and he shared the insights generated from a fourth AI Lab participant – The Durango Herald. More impacts from the AI Lab’s efforts will be available soon as the other lab participants are concluding their experiments.
The Durango Herald, Ballantine Communications
Herald the Helper, a chatbot to answer readers’ questions
The question The Durango Herald in Colorado and its Managing Editor Shane Benjamin wanted to explore was: Can a chatbot with AI capabilities provide readers with useful information while providing journalists with insights into what readers are interested in or needing more information about?
According to Benjamin, they first decided to give it (their chatbot) a personality — and a local twist. “Harold,” a friendly Sasquatch, lives on the Herald’s website, helping readers submit tips and find stories from the paper’s archives, according to Benjamin.
Within minutes of going live, Harold received a message from a reader about a child injured in a chairlift accident, something the newsroom hadn’t yet heard about. The newsroom immediately made phone calls and confirmed the story. Harold helped them break news, Benjamin said.
***
SE Missourian, Rust Communications
AI editorial assistant for reporters and editors
The question Rust Communications President Jon Rust wanted to explore was: Can AI be a powerful tool to improve the quality of journalism – including questions asked and answered in stories – in a more efficient and effective way for news operations than traditional copy editing, especially at a time when copy editing is becoming less prevalent? In addition, can such a tool also be used to train writers in journalism best practices and ethics, expanding the pool of talent for news operations in rural and/or other underserved areas.
The results were striking: 79% of reporters and 89% of editors said the AI analysis improved story quality, according to Rust. Some reporters began using the system earlier in the writing process to prepare for interviews and identify follow-up angles.
***
The Baltimore Times
Community News Initiative
The question The Baltimore Times and Publisher Paris Brown wanted to explore was: How can AI be a tool to help a news media company increase content creation, enhance community engagement, and grow The Baltimore Times audience?
Brown said that through its “Community Newsroom Initiative,” the Times is testing an AI-driven system that allows readers and local organizations to submit stories via a standardized form. The AI helps shape submissions into publishable drafts in the paper’s editorial voice —always with human review before publication.
The intent is to produce more content, expand audience — especially among younger readers — promote sustainability, grow The Times’ digital reach, encourage greater community engagement, and improve editorial workflows, Brown told the LMA Fest audience.
***
Shaw Media
Transforming written content to audio content
The question Shaw Media and Managing Editor John Sahly wanted to explore was: Can AI be successfully implemented in a way that the audience accepts, thereby building community trust and keeping the audience informed? Shaw experimented with converting text to audio.
Sahly told the LMA Fest audience that early listener feedback has been enthusiastic, especially from former residents who use the feature to stay connected to their hometowns.
One listener wrote that hearing local stories again made them “feel closer to the community they left behind.” Sahly said that captures why this (AI Lab) experiment matters: It builds trust by meeting audiences where they are — on their phones, in their cars or through a smart speaker.
Highlights from the Q&A with the panel
Editor’s Note: Responses are heavily cut from the original transcription to communicate their primary points.
What was the biggest aha moment during the experiment?
John Sahly: The team recently launched AI-generated playlists featuring the five most-read stories in six different counties. In just two weeks, reader feedback has been positive — an early sign of success and a potential step toward using AI to strengthen trust by meeting audiences where they are.
Paris Brown: I was approached by someone, and they told me that we didn’t have capacity. “So I thought, how can I create it?”
Jon Rust: Reporters got constructive feedback that they could look at themselves — see it, digest it, think about it and incorporate it. For our young reporters especially, this feedback was less intimidating than a human editor. They really gravitated towards it.
***
What do you think the future of AI in news will look like a year from now?
Jon Rust: We just keep experimenting. But one-to-one story versioning for preferences of the reader will be one thing; for example, some people prefer a condensed story when an article goes over a certain length. We see this as an accessibility issue.
Paris Brown: A CMS system that’s AI-powered, that takes you from tasks to podcasts to video to audio with a social media-first approach.
John Sahly: It’s whatever we make of it. And by that, I mean the barriers to entry have never been fewer. It’s never been cheaper. It’s going to be whatever you make of it.
***
How do you ensure that a human is involved when AI tools are used?
John Sahly: I think a human will always have to be in the loop.
Paris Brown: Absolutely, I think humans need to be involved. I mean, you’re the driver. You’re the prompt engineer.
Jon Rust: We’re very clear that when it has to do with story generation or augmentation, all of that has to have human oversight. We have it in our company policy. We talk about that at our company meetings.
***
Editor’s note: AI was used to transcribe the audio from this session at LMA Fest. It was then used to create the first draft of this story, which was then edited by Local Media Association.
You must be logged in to post a comment.