AI-generated cartoons and columns expose why newsrooms need stronger standards and transparency

When editors let writers turn to chatbots to do their writing, they’re also outsourcing the thinking and creativity they sought in the first place.”

Rob Tornoe, digital editor and writer for The Philadelphia Inquirer and E&P cartoonist and columnist

Tech leaders are suddenly on the same page in suggesting a pause in the development of artificial intelligence (AI). It might be time to consider a similar pause in the newsroom. 

As the hype machine for AI overtakes most reasonable uses of the technology (while also threatening to maybe, possibly kill us all), we as an industry need to take a second to determine how we envision using the tools to benefit our readers and the larger ecosystem of journalism. 

I come into this with my own shade. As a writer, I see the slop pumped out by robots across the internet and worry about news organizations — especially those owned by hedge funds — driving away our declining readers in the search for more profits. As a cartoonist, I abhor how images can be created at the click of a button by tools largely built by stealing the work of real working artists. 

There are certainly things AI-powered tools are good at, but producing original content isn’t among them. That brings us to the Baltimore Sun and the need for consistent AI standards in the newsroom. 

In August, staffers turned to AI-powered tools to create a cartoon for the opinion page, a mangled mess hyping up Mayor Brandon Scott that embodies the term “AI slop.” 

Veteran editorial cartoonist Scott Stantis, who currently draws for the Chicago Tribune, said Baltimore Sun opinion editor Philip Caldwell reached out to him in early August about drawing local cartoons. Stantis was interested, but said Caldwell later replied with a sadly common response — there was no money in their budget for editorial cartoons. 

A week later, a local cartoon hyping up Mayor Brandon Scott appeared with the caption, “Produced using Artificial Intelligence tools by the Baltimore Sun Editorial staff.” 

“It just sucks,” Stantis said on his “DMZ America Podcast” to co-host Ted Rall and fellow cartoonist Rick McKee. “You know the meeting where they did this … They sat around the table going, ‘Damn, this is good.’” 

Stantis is probably right, because the Baltimore Sun has continued to publish AI-generated cartoons, despite near-universal criticism from editorial cartoonists across the country. Making matters worse is the Baltimore Sun once ran work drawn by Pulitzer Prize finalist Kevin Kallaugher, who was dismissed last year, he contended, because his politics differed from the paper’s owner, Sinclair Broadcast Group Chairman David D. Smith. 

“Whatever the intent, the only message the Sun actually delivered was that it has finally and completely surrendered any remaining claim it may have had to journalistic legitimacy and its former place among the nation’s important newspapers,” Marc Murphy, the president of the Association of American Editorial Cartoonists, wrote in an open letter. 

The Herb Block Foundation went further, calling the work “borderline immoral.”

“This is a textbook example of the threat AI poses to human jobs,” wrote cartoonist Rob Rogers, a board member at the foundation. “AI companies scrape artists' work from the internet without permission, credit, or compensation to train generative models. These corporations make billions at the expense of others.”

Neither Caldwell nor Baltimore Sun Publisher and Editor-in-Chief Trif Alatzas responded to requests for comment. 

The use of AI-powered tools on the opinion page hadn’t been limited to lifeless attempts at generating cartoons. 

In late August, The Wall Street Journal defended publishing an opinion piece supposedly written by billionaire investor Stanley Druckenmiller criticizing Treasury Secretary Scott Bessent. But after some internet sleuthing, Druckenmiller admitted turning to an AI-powered large language model to write his piece, telling NOTUS (now rebranded as the Washington Sun), “I write everything using AI now.” 

Editorial page editor Paul Gigot responded by mocking Poynter’s finger-wagging and slamming the larger media world over how little we’ve “thought about the reality of AI” and the larger role of the opinion page.

“Many contributors, especially prominent politicians or CEOs, employ speechwriters or hire a Beltway writing shop,” Gigot wrote. “How different is that from using AI to streamline an essay?”

I would say maybe we should question the policy of allowing columns to be ghostwritten as well, but that’s getting off the subject. When editors let writers turn to chatbots to do their writing, they’re also outsourcing the thinking and creativity they sought in the first place. It also comes down to trust, and right now readers trust human beings over robot-written slop with a tendency to make up facts and pump out annoyingly obtuse copy. 

Also in August, the Financial Times was forced to add an awkward note to a column written by Harvard professor Ricardo Hausmann, who used AI tools to condense his piece before he submitted it to editors. Unlike The Wall Street Journal, the Financial Times “specifically prohibits the use of AI in the writing process.” 

I don’t know how I feel about the cottage industry cropping up in media circles trying to out the use of AI-powered tools in journalism, but it’s certainly a growing issue. Semafor’s Max Tani reported 10 out of 310 guest columns published by The Wall Street Journal, The Washington Post and The New York Times were labeled at least 80% AI generated, according to AI detector Panagram. 

If anything, I was surprised the number was so low. While Gigot made it clear The Wall Street Journal will continue to have a stricter AI policy for staffers than contributors, both The New York Times and The Washington Post prohibit guest columnists from using AI to help draft and develop their essays, and didn’t respond to Semafor about appearing to run afoul of those guidelines. 

For what it’s worth, most editors seem reluctant to follow Gigot’s path, at least for now. Nieman Journalism Lab’s Joshua Benton spoke with 24 editors at prominent newspapers and said none were willing to embrace a robot-written approach. 

“Our audience is paying for the words and thoughts of human beings,” the San Francisco Chronicle’s Matthew Fleischer told Benton. 

So, what can newsrooms do? For starters, develop a specific code of conduct involving the use of AI-powered tools and apply it to all editorial content, including those submitted by contributors not on staff. 

“An AI policy pinned to a newsroom Slack channel doesn’t do much good when the person submitting the copy works somewhere else,” Poynter’s Ren LaForme recently wrote.

It gets trickier when editors are turning to syndicated columns for their op-ed pages. 

None of the major content syndication companies appear to have a public-facing AI policy regarding their own content. Tribune Content Agency’s AI policy guidelines are mostly focused on restricting how subscribers can use their content. So, it may come down to editors needing to scan pieces before publishing them on their own platforms. 

Regardless, readers need clarity and transparency when it comes to content generated using AI tools. And maybe a few human beings left in the process. 

Rob Tornoe is a cartoonist and columnist for Editor and Publisher, where he writes about trends in digital media. He is also a digital editor and writer for The Philadelphia Inquirer. You can reach him at robtornoe@gmail.com.