Last week, Baltimore Sun reporters were shocked to see in their publication two stories produced not by them or by some non-staff journalist or by any human being at all – but by AI.

One piece was an analysis of Maryland Gov. Wes Moore’s State of the State speech. The other analyzed “the public exchange” that followed Moore’s speech, including social media posts by President Donald Trump.

Instead of a byline, a note at the top said these “analyses” were “generated by an artificial intelligence tool at the request of the Baltimore Sun and reviewed by staff members.”

“AI slop,” union leaders quickly branded them, using the now commonly accepted pejorative term for digital content made with generative artificial intelligence that is lacking in effort, quality or meaning.

“Sun management has once again disparaged our talented HUMAN reporters and their work, this time by filling more than half a page in today’s paper with AI slop,” the Baltimore Sun Guild tweeted.

The union also pointed out a glaring error: one of the pieces twice referred to Trump as “former President Donald Trump.”

“How thorough was that “review” of the AI slop, anyway?” the union said in another post.

The reaction on Reddit to the stories, which appeared in the print paper on Feb. 13, was even more scathing.

“This is gross and pointless, except as a test of an AI capability so they can continue to cut staff and make all creative jobs either completely over or greatly reduced as the humans will just be serving the AI while it does 95% of the work,” one commenter wrote.

“What a DISGUSTING comment. It’s like you haven’t considered the shareholders at all!” another responded mockingly. “My butler is going to performatively laugh SO hard when I tell him about this comment.”

Management: It’s Innovation

Sun management, meanwhile, though chastened by the mistakes in the copy, defended the idea of using AI for editorial content as innovation and vowed to continue exploring it.

Publisher and editor-in-chief  Trif Alatzas responded to The Brew’s query with this emailed statement:

“We thought this method provided an opportunity to try an alternative way to analyze these subjects.

Our staff reviewed the work and discussed how to ensure we were transparent with readers. One reader said he appreciated The Sun embracing such innovation and encouraged us to try it more; he added we must thoroughly review future efforts given the error we should have caught and have since corrected. Another reader told us to knock it off.

We plan to work with our staff to find various ways to use the technology responsibly to aid our reporting and better serve readers. There might be some mistakes along the way, but we plan to learn from them and be transparent about our efforts.

To ignore this powerful tool would be a mistake.”

Alatzas did not respond to questions about why the staff, including the news organization’s political reporters in Annapolis who cover Moore, were not informed about the use of AI for material on their beat.

Guild: “Disparaging reporters”

Reporter Dan Belson, the Baltimore Sun Guild’s acting unit chair, said he heard about the AI “stories” from a former Sun staffer who happened to see them.

The staff, he said, found management’s decision to begin publishing AI-generated content side-by-side with their professionally-produced journalistic work demoralizing.

“We would not like to see our work that we worked very hard on being replaced by the substandard content that we saw on Friday,” he said.  “It’s embarrassing to see that as a substitute for what we do, it’s disparaging our reporters.”

“For them to go forward with investing in something like this – by this ‘content generator’ that doesn’t talk back – it’s insulting,” he said.

“It’s embarrassing to see that as a substitute for what we do, it’s disparaging our reporters”  – Dan Belson, Baltimore Sun Guild.

Belson said the AI content was especially irritating given that it comes after two years of stalled contract negotiations with management that began shortly after new owners acquired the newspaper.

Sun majority owner David D. Smith, Sinclair Broadcast Group Executive Chairman, made no effort to disguise his low opinion of the publication in his first meeting with its staff in January 2024, noting that he “hadn’t read the newspaper in 40 years.”

Since then, there have been frequent clashes with the Guild over management’s impact on journalistic quality, with members denouncing it as unprofessional and biased, leading to streetside picketing and the firing of one reporter after she raised concerns about the issue internally.

Belson said the use of AI in this instance was especially troubling because it was employed to report on a topic that is so central to the mission of the press.

“It’s about politics, it’s about elected officials, it’s about government,” he said. “It’s not just a ‘Hey we asked AI how the Ravens are doing’ kind of thing.”

Asked if the pieces were intended by management to be anti-Wes Moore and pro-Donald Trump, Belson said he didn’t know, but called them failed efforts at accountability reporting.

“They’re written by something that does not have the capability of logic,” he said. “The piece about Trump’s response to Wes Moore doesn’t even include the Truth Social post that it’s talking about – the piece is just hard to read.”

The other article “criticizes Wes Moore for not saying, like, exactly when the homicides dropped by 50%, but we have that data, it’s a known fact,” Belson continued. “We’ve reported extensively on the decline in homicides – a human could figure that out.”

Belson said the union plans to formally reach out to management tomorrow to demand transparency about their plans for AI and to seek a pledge to adopt standards for its use such as the ones published as part of the national NewsGuild/Communication Workers of America newsnotslop.org campaign: