■ HIGH RISK ■ Media & Communication
AI can already turn press releases, box scores, and meeting agendas into passable copy, which automates a big slice of what shrunken newsrooms still produce. But the local reporter's economic problem is the business model, not the robot — and the work that matters, showing up and finding out what officials aren't announcing, remains stubbornly human.
“AI writes local news stories from police blotters. Your shoe-leather journalism is romantic but slow.”
Our AI replacement risk score — how we score jobs
Local reporting is a volume business run on too few people: covering council and school-board meetings, working police and court records, writing up business openings and Friday's game, cultivating sources at city hall, filing public-records requests, and — when time allows, which is rarely — the accountability work of finding out what the agenda didn't mention. One reporter often files multiple stories a day across beats that used to employ five people, because the industry lost most of its ad revenue to platforms long before AI showed up.
Automation slots neatly into the volume end. Template systems have written earnings recaps and sports briefs for years; LLMs now competently draft stories from press releases, police blotters, agendas, and meeting transcripts, and AI transcription makes every recorded meeting searchable. Some chains openly use AI to expand cheap coverage, and 'pink slime' outlets generate algorithmic local-ish content with no reporters at all. For publishers surviving on thin margins, the temptation to let the model write the rewritten-press-release tier of content is irresistible — and that tier is a large share of what remains of local news.
The resistant core is presence and trust. A model cannot notice the councilman flinch, develop a source in the clerk's office who slips you the real numbers, knock on the door of a grieving family and be let in, or sit through the executive session's aftermath asking who benefits. Original reporting generates the facts everything downstream summarizes; AI only rearranges what someone already found out. The bitter irony is that the endangered part of the job pays the bills and the essential part doesn't — so the realistic future is fewer generalist reporters, AI producing the routine layer, and human journalists concentrated on enterprise and accountability work funded increasingly by nonprofits and memberships.
Automatability: our editorial assessment of current and near-term AI capability
The routine-copy layer is automating right now — chains are already using AI for briefs and recaps, and it spreads through the late 2020s. Since local newsrooms were shrinking for economic reasons anyway, AI accelerates an existing decline rather than starting one. Reporters doing original, source-driven work remain irreplaceable but scarce, with nonprofit and member-funded outlets becoming their main habitat by ~2030.
It's accelerating a decline that platform advertising started two decades ago. AI now writes the routine layer — briefs, recaps, rewritten releases — cheaper than junior reporters, and low-quality algorithmic 'local' sites fill the vacuum where papers died. What AI cannot do is original reporting: developing sources, obtaining records, witnessing events. The crisis is that the market pays badly for exactly that.
It can transcribe the livestream and summarize the agenda items accurately, which honestly covers what overstretched outlets often published anyway. What it misses is everything off-script: the argument in the hallway, the item pulled without explanation, the contractor's name that keeps recurring. Meeting summaries inform; reporters who notice anomalies and chase them are how communities find out what's actually happening.
Only with clear eyes and a specialty. Generalist churnalism jobs are evaporating between economics and automation. But investigative, court, and beat reporters with source networks and records skills remain scarce and valued — increasingly at nonprofit and digital-native outlets rather than legacy papers. Enter planning to do the work AI can't, and build a portable audience from day one.
Use it as a force multiplier for grunt work: transcribing interviews, summarizing documents, drafting the routine brief so your afternoon goes to the enterprise story. The reporters who thrive will file the same volume of routine copy in a fraction of the time and reinvest the hours in original work — the stories that justify a subscription and a byline.