How AI Is Being Used to Paper Over a Local News Crisis
McClatchy’s Content Scaling Agent is publishing AI-written stories under reporters’ bylines, and while traffic numbers look promising, the subscriber figures tell a different story.
McClatchy, one of the U.S.’s leading local and regional news publishers, is using AI-generated content as part of its attempt to address a $34M drop in revenue, weaker audience growth and pressure on newsroom productivity, according to a report by Straight Arrow’s Mark Keierleber.
Keierleber obtained an exclusive recording of a McClatchy town hall in which executives told journalists that AI adoption was now central to the company’s future despite serious misgivings by staff journalists.
The publisher has already built an in-house ‘content scaling agent’ which puts the bylines of real reporters onto AI-generated stories and, in April, wrote four of the company’s top five articles. AI use has also been incorporated into job descriptions and performance reviews.
For publishers, the significance is that AI is being viewed as a business survival tool and automated output is now seen as a proxy for audience value, when in reality the underlying problem is weakening reader habit, a cost of living crisis, and a perceived lack of utility.
The numbers explain McClatchy’s urgency
In an interview with WNIP, Keierleber says the most striking part of the McClatchy story is the extent to which executives appear to see AI as a way to address commercial pressure.
“I was most surprised by the degree to which McClatchy executives view artificial intelligence as a key to solving their business woes — and not necessarily as a way to improve the quality of their work,” he said. “They clearly see AI as a way to generate a bunch of articles with the hope that some of them will resonate.”
According to Keierleber, McClatchy executives told staff that revenue was down $34 million year-on-year. The company had also missed its own pageview targets, while its sites were seeing fewer than 4 million new known users per month, down from 15 million a year earlier. Most concerning, he said, was that McClatchy had begun to see a decline in digital subscribers.
If a publisher is losing subscribers while generating millions of visits from automated articles, the commercial signal is awry. Traffic may provide short-term advertising inventory, but it does not necessarily create the habit, trust or perceived usefulness that supports reader revenue.
Keierleber says, “The McClatchy numbers suggest the AI-generated stories are generating cheap clicks but may not contribute to the reader trust or loyalty critical to news business sustainability.”
The American Press Institute’s Yoni Greenbaum, writing on his Substack, succinctly sums it up, “AI isn’t going to replace journalism. It’s going to replace what too many news organisations have been calling journalism: the commodity content, the wire rewrites, the SEO plays, the press release transforms.”
Greenbaum adds, “Loyal subscribers describe their news brand as the lens through which they interpret the world — a cognitive and emotional home base, not a content vendor.”
AI can improve workflows, but where’s the strategy?
McClatchy defends itself by saying it has 35 more reporter positions across its brands than it had a year ago. But Keierleber’s example from the Centre Daily Times illustrates the problem.
A reporter had written about renovations at the Pennsylvania Military Museum, making clear that the site would not reopen until the following year. The AI system later repurposed the story for Memorial Day weekend searches, producing a piece that told readers hoping to visit that the doors were still shut. It added little new information, was pockmarked by AI-generated bullet points, and looked designed to capture search demand rather than solve a genuine reader need.
The most sensitive part of McClatchy’s approach is the use of reporter bylines on AI-generated articles. For readers, a byline signals accountability but when AI-generated material appears under that name, the reader may not understand where human reporting ends and AI begins. If an AI hallucinates, the problem becomes serious as the individual reporter’s credibility absorbs the damage.
Yet the system, incentives and rollout are being set at executive level.
Keierleber says, “Much of the debate over AI use in newsrooms has centered on the narrative of the lazy journalist,” he said. “In fact, much of the push for AI adoption in newsrooms has come at the executive level in the name of efficiency.”
Indeed, several McClatchy journalists told Keierleber that the Content Scaling Agent has added to their workloads rather than reducing them, requiring them to check and correct AI output on top of their existing responsibilities. The efficiency argument, central to the executive case for adoption, is not bearing out in newsroom reality.
McClatchy is not an outlier
The Associated Press, USA Today, and dozens of other media outlets across the industry are integrating AI into their content operations to varying degrees, and why shouldn’t publishers use it where it reduces low-value work, improves reporting capability or helps journalists serve readers better?
But the McClatchy numbers suggest a serious misjudgment: AI is generating clicks without generating loyalty, and declining subscriber figures indicate that volume is no substitute for value. Basically, AI-generated content could be making the old scale problem cheaper to repeat.