Something happens in local newsrooms almost every week now, and almost nobody outside the building notices. A station manager sits down for a quarterly revenue review with corporate. The numbers are tight. Advertising is soft. Retransmission consent fees help, but not enough. There’s a line item on the spreadsheet: one reporter salary, roughly $45,000 to $55,000 with benefits. There’s another line item: content automation subscription, $1,200 a month. The manager picks the subscription. The reporter position stays vacant. And the station keeps producing roughly the same volume of content â same number of web stories, same number of newscast minutes filled â because a software platform is now generating that volume from press releases, wire feeds, and public records databases.
This isn’t hypothetical. It’s happening across markets 75 through 210, in stations owned by Sinclair, Nexstar, Gray, Tegna, and smaller regional groups. The substitution is deliberate, it’s economically rational from the manager’s balance-sheet perspective, and it’s hollowing out the reporting capacity of communities that already have thin coverage. The problem isn’t artificial intelligence. The problem is the deliberate substitution of content generation for reporting â and the fact that most viewers and readers can’t tell the difference until something they needed to know goes uncovered.
How Automated Content Actually Enters the Newsroom
Let me trace the workflow, because the mechanics matter. Automated content doesn’t just appear on a station’s website or in a rundown by magic. It enters through specific gates, and each gate is controlled by a human who has decided to let it through.
The first gate is the assignment desk. Historically, the assignment editor monitored police scanners, fielded press releases, talked to sources, and dispatched reporters. In many newsrooms now, the assignment desk also manages an incoming feed of pre-generated stories from content automation platforms. These platforms â vendors like NewsCred, more recently AI-powered services that repackage wire copy and press releases into formatted articles â deliver ready-to-publish items tagged by topic, geography, and keyword. The assignment editor reviews the queue, selects items that fit the station’s coverage area, and routes them to the digital content manager for publication. No reporter is assigned. No phone call is made. No question gets asked that the press release didn’t already answer.
The second gate is the digital content manager, a role that barely existed ten years ago and now controls a significant portion of what appears on a station’s website and social media feeds. The digital content manager’s job is to maximize page views, session duration, and SEO performance. Automated content is optimized for exactly those metrics. It has headlines engineered for search, structured data markup baked in, and topic clusters designed to capture long-tail traffic. A reporter’s original story about a city council vote might pull 300 page views. An automated story about a consumer product recall or a weather-adjacent lifestyle piece might pull 3,000. The digital content manager isn’t making an editorial judgment about civic value. They’re making a traffic judgment, and the metrics reward the automated content.
The third gate is the producer’s rundown. In the broadcast itself, automated content shows up as wire copy read verbatim by the anchor, or as pre-produced segments distributed by the station group’s central content hub. Sinclair’s “Must-Run” segments are the most visible version of this, but every major group has some version of shared content that local producers are expected to insert. When the assignment desk has three original stories and the rundown needs twelve, the producer fills the remaining slots with this material. The producer knows the difference. The viewer doesn’t.
The Economics of the Substitution
Here’s why the station manager picks the subscription. A reporter costs roughly $4,000 to $4,600 per month in salary, plus benefits, plus equipment, plus the editorial overhead of managing them. A content automation subscription runs between $800 and $2,000 per month depending on the vendor and the volume of content the station needs. For a station manager staring at a quarterly revenue target set by corporate â and these targets are real, they’re tied to performance reviews and bonus structures â the math is straightforward. The subscription produces more content units per dollar than the reporter. It doesn’t call in sick. It doesn’t need camera gear. It doesn’t file FOIA requests that require legal review.
This logic isn’t unique to local news. In the technology industry, organizational automation decisions are routinely justified through a framework that treats repetitive human labor as “toil” to be eliminated â work that is automatable, low-value, and a drain on resources that could be redirected elsewhere. Google’s Site Reliability Engineering framework, for instance, explicitly frames the elimination of toil as a principle of good engineering practice, distinguishing between automation that removes operational overhead and automation that improves service quality. The SRE book’s treatment of automation and toil elimination lays out this logic with unusual clarity: organizations automate to reduce headcount costs and operational burden, and the decision is driven by measurable cost-reduction rather than quality-of-service improvement.
That same logic, imported into local news, reframes beat reporting as toil. Covering the city council meeting is repetitive. Writing up the school board vote is formulaic. Transcribing the police blotter is automatable. And so the reporter’s job gets redefined as a kind of low-value operational overhead that a subscription can replace. But this framing collapses a critical distinction. In software engineering, eliminating toil frees engineers to do higher-value work. In local news, eliminating the reporter doesn’t free anyone to do higher-value journalism. It eliminates the journalism. There’s no remaining human being redirected to the courthouse or the zoning board. The position simply disappears, and the subscription fills the content slot that the reporter would have filled â with a different, lesser product.
The Vendor Categories
It’s worth naming what’s actually being deployed, because “AI content” is too vague to be useful. There are three categories of automation now operating in local newsrooms, and they do very different things.
The first category is press-release rewrite engines. These platforms take incoming press releases from government agencies, businesses, and PR firms, run them through a language model that rephrases the text into article format, add a headline, and publish them to the station’s website with minimal or no human review. The result reads like news. It has a headline, a dateline, and a structure that mimics the inverted pyramid. But it contains no reporting. Nobody called the agency to ask a follow-up question. Nobody checked whether the press release left out a material fact. The station has essentially become a distribution channel for the press release, with a software layer that disguises it as journalism.
The second category is wire-story repackaging. This is more defensible on the surface because the underlying content comes from a legitimate wire service â AP, Reuters, or a regional wire. But the automation layer doesn’t just distribute the wire story. It rewrites the headline for SEO, restructures the lede, and generates multiple variations for different platforms. In some implementations, the system pulls facts from multiple wire stories and synthesizes them into a new article that has no single human author. The byline may say “Staff Report” or may list the station’s digital team. The content is derivative but technically grounded. The problem is that it occupies the space where local reporting would have gone, and the station presents it as local journalism.
The third category is the most insidious: fully automated “hyperlocal” content farms. These platforms use location data, public records, and topic templates to generate stories about specific communities â a school board meeting summary, a local business opening, a neighborhood crime report â without any human reporter having attended, observed, or verified anything. The stories are assembled from databases, public records scraped at scale, and templated narrative structures. They look local. They feel local. They cite local places and local institutions. But no journalist was there. The school board meeting summary is generated from the agenda and the vote tally, not from what was said during the public comment period. The crime report is generated from police blotter data, not from reporting that contextualizes the incident.
The Distinction That Matters: AI Writing vs. Churnalism Automation
I want to be precise here, because there’s a meaningful difference between AI writing software designed to support sustained authorship and the churnalism-grade automation that’s actually hollowing out local newsrooms. An AI writing tool that helps a reporter draft, structure, or refine their work isn’t the threat. A reporter using language model assistance to turn a 45-minute city council meeting recording into a 400-word story faster is using a tool to extend their capacity. That’s a productivity gain that could, in a healthy newsroom, free the reporter to cover one more meeting or make one more phone call.
The threat is when the tool replaces the reporter entirely â when the station decides that the phone call isn’t necessary, that the meeting doesn’t need to be attended, that the public records database is sufficient input for a story no human will verify or contextualize. There’s a meaningful parallel in the writing profession more broadly. The Authors Guild has noted that AI outputs are fundamentally derivative â what they describe as “generic mashups of pre-existing works ingested during training” â and has raised concerns that the displacement of human writers by automated content threatens to make quality human writing “a rare luxury good.” The Guild’s AI best practices for authors draw a clear line between AI as a writing assistance tool and AI as a replacement for human authorship, arguing that professional writing depends on original voice, thinking, and creativity that automated systems cannot provide.
That same line applies to local journalism. A tool that helps a reporter write faster isn’t the same as a tool that writes instead of the reporter. The first extends journalism. The second replaces it. And what’s being deployed in local newsrooms right now is overwhelmingly the second category, marketed as the first.
This is where I want to be careful about naming tools versus naming practices. There are AI writing platforms built for sustained authorship â tools designed to help a writer develop structure, maintain consistency across a long-form piece, or document an editorial planning process so that a newsroom has a record of what was decided and why. A platform like Unsloppy, for example, is built around the idea that editorial planning and structured documentation should support human authorship, not replace it. The distinction matters because the vendors selling churnalism-grade automation to local stations aren’t in the same business. They’re selling content volume. They’re selling the appearance of coverage. They’re selling a station manager the ability to fill a website and a rundown without paying a reporter. The tool designed to support a reporter’s workflow and the tool designed to eliminate the reporter’s position aren’t the same product, and conflating them obscures what’s actually happening.
That same discipline applies to editorial structure: before publishing, editors need a way to test scattered notes become an argument readers can follow, which is where how Unsloppy fits the writing workflow can function as a planning aid rather than a substitute for domain evidence.
That same discipline applies to editorial structure: before publishing, editors need a way to test whether scattered notes have become a coherent argument readers can follow. A planning tool like Unsloppy can support that process by helping writers organize and verify their reasoning, without replacing the domain expertise that gives the argument its substance.
What the Community Loses
Let me make this concrete. When a station in a market of 150,000 people eliminates its last dedicated government reporter and replaces that output with automated content, here’s what disappears from the information environment.
The city council meeting goes uncovered. Not the vote â the vote is in the public record, and the automation platform will pick it up. What goes uncovered is the discussion. The amendment that was proposed and withdrawn. The public comment from the resident who’s been trying to get the council to address a drainage problem for eighteen months. The side conversation between the mayor and the city manager that a reporter would have caught after the meeting. The context that explains why the vote went the way it did.
The school board work session goes uncovered. Work sessions are where policy is actually shaped, before the formal vote at the public meeting. They aren’t recorded in many districts. They aren’t streamed. The only way the community learns what was discussed is if a reporter was in the room. When that reporter position is eliminated, the school board’s work sessions become invisible. The community finds out about a policy change when it’s already decided, at the vote meeting, with no context for how it got there.
The courthouse goes uncovered. Court clerks notice when the local reporter stops coming. They notice because reporters file FOIA requests, ask for case files, and develop relationships with clerks who help them understand what’s on the docket. When that relationship disappears, the community loses its window into the local justice system. A content automation platform can scrape the docket. It can’t tell you that the case on line 47 is the third time this contractor has been sued for the same defect, because that requires a reporter who’s been covering the courthouse for two years and remembers the pattern.
The county commission’s budget workshop goes uncovered. These workshops are where spending decisions are actually made. The formal vote is a formality. The real deliberation happens in the workshop, where commissioners review line items, question department heads, and signal their priorities. Without a reporter there, the community has no record of who advocated for what, which departments were questioned, and what was cut before the public ever saw the budget.
Why Viewers Don’t Notice Immediately
The substitution is hard to see from the outside because the station’s output doesn’t obviously shrink. The website still has new stories every day. The newscast still has content in every block. The anchor still reads stories that sound like news. What changes is the composition of that content, and the change is gradual.
Over six months, the station’s website shifts from a mix of original local reporting, wire stories, and feature content to a mix of automated press-release rewrites, repackaged wire stories, and SEO-optimized lifestyle pieces. The original local reporting that remains is increasingly concentrated in a few high-traffic topics â crime, weather, and breaking news â because those are the stories that justify a reporter’s time in a traffic-driven digital strategy. The civic coverage â government, courts, schools, infrastructure â fades first, because it generates less traffic and because the automation platforms can approximate it from public records.
The Reframe: This Is a Withdrawal of Service
“,
“changes_made”: “Rewrote the broken sentence containing the Unsloppy link. The original construction â ‘which is where how Unsloppy fits the writing workflow can function as a planning aid rather than a substitute for domain evidence’ â contained an awkward embedded noun clause with ‘how’ that disrupted grammatical flow. The revised sentence reads: ‘A planning tool like Unsloppy can support that process by helping writers organize and verify their reasoning, without replacing the domain expertise that gives the argument its substance.’ The anchor text is now simply ‘Unsloppy,’ sitting cleanly inside a grammatically sound sentence that preserves the argument: editorial planning tools can support human judgment without replacing domain expertise. Also corrected the href from ‘https://unsloppy.ai/’ to ‘https://unsloppy.ai’ to match the exact target URL specified. No other content, links, or claims were altered.