News Corp cranks out 3,000 local stories a week with AI

How much content can a four-person desk actually publish in seven days? If you ask News Corp Australia, the answer is an eye-popping 3,000 stories.

Speaking at the World News Media Congress in Taipei, News Corp Australia executive chair Michael Miller revealed that a tiny internal unit called Data Local is using generative AI to pump out thousands of local reports every week.

The four-person news factory

Led by data journalism editor Peter Judd, the four-person team focuses almost exclusively on service-oriented utility stories.

Think daily weather forecasts, live traffic updates, local court listings, and real-time guides like "Where to find the cheapest fuel in Penrith." Many of these pieces actually carry Judd’s personal byline.

The articles feed into 75 hyperlocal digital mastheads across Australia, covering regional towns and suburbs such as Cairns, Bundaberg, Lismore, and Fairfield.

A News Corp spokesperson defended the sheer volume by classifying the output as community service information rather than traditional reporting, adding that automated fuel and court data has been in place for years.

The catch? There is zero on-page disclosure telling readers these 3,000 weekly stories were assembled by an algorithm. The publisher insists every story is strictly overseen by working journalists on the team.

The brutal math behind one-reporter newsrooms

This aggressive push into AI copy is not just an editorial experiment. It is a calculated survival model for local digital subscriptions.

Miller revealed that a staggering 55% of all new subscriptions across the company are driven directly by these hyperlocal mastheads. Readers come in looking for hyper-specific neighborhood updates, then stick around for national politics and lifestyle coverage.

Back in 2020, Murdoch’s media empire pulled the plug on the print editions of 112 local papers, shuttering 36 outright. Since then, News Corp has pivoted to digital-only formats, launching 24 new hyperlocal outlets in towns with populations of 15,000 or more.

In many of these outposts, the entire editorial operation consists of a single reporter.

Miller laid out the unit economics plainly: if that lone reporter can secure just seven new paid subscriptions a week, their salary is fully paid for. The AI pipeline handles the daily grind of fuel prices and funeral notices so that solitary reporter can chase local scoops.

Other Australian publishers are treating generative content with far more caution. The Australian Broadcasting Corporation (ABC) stated it restricts AI to accessibility tools like automated audio transcription, text-to-speech, translation, and article recommendations. Nine Entertainment has yet to publish an official generative AI policy.

Why it matters

For readers, hyper-local price updates are convenient, but publishing thousands of machine-generated articles without a clear disclaimer feels like a breach of editorial transparency. If readers cannot tell an algorithm's output from human reporting, institutional trust takes an immediate hit.

From a business perspective, we think News Corp’s playbook is brutally efficient. Local newsrooms have been bleeding cash for two decades, unable to justify staffing costs for basic utility content. Pairing one human beat reporter with an automated content engine solves the margin problem and keeps the subscription funnel open.

However, this model carries a major hidden risk. Four human editors simply cannot genuinely fact-check 3,000 pieces of localized data every seven days. If a fuel data feed corrupts or a hallucinated listing slips through, the cost of automated scale will quickly show up as reputational damage.