AI-powered automated news service Radar now writes thousands of stories a month, including general interest stories, for up to 350 local news titles in the UK
Patricia Nilsson / Financial Times : Tweets: @financialtimes and @matthausk Tweets: @financialtimes : 'The readers couldn't care less who wrote it, frankly.' https://www.ft.com/... Matthus Krzykowski / @matthausk : Imo people underestimate how far along we already are in audio/text/image content automation and generally do not understand that related challenges won't go away by policing platforms such as FB, Google or Twitter https://twitter.com/...
Context & Ripple Effects
Radar's move into general-interest stories is the latest step in a decade-long march that began with templated sports and earnings copy — the lineage runs from Vista's acquisition of Automated Insights, the startup behind the AP's robot news writing, through Bloomberg News automating around a third of its output. What changes here is scale and scope: thousands of stories a month spread across up to 350 UK local titles, no longer confined to data-shaped beats.
The timing matters because the economics are converging from both sides — publishers cutting costs and generators getting cheaper. The related coverage on [[a:839755|AI text quietly authoring more of the internet, with fewer clients buying human-written content]] frames Radar not as an experiment but as a working commercial model for local news.
First-order effects
- Reporters and subs at those up to 350 UK local titles shift toward editing, sourcing, and oversight as routine story volume is machine-produced — the FT quote in the piece ('readers couldn't care less who wrote it') signals bylines are losing their gatekeeping role.
- Radar's publisher clients gain volume at near-zero marginal cost per story, letting thin local desks cover general-interest ground they previously dropped.
Second-order effects
- Freelancers and agencies selling commodity local coverage face a collapsing price floor, accelerating the client attrition the Washington Post reporting describes.
- Competing local news groups must either license similar automation or differentiate on original reporting — the same fork the BBC faced when it chose to build its own AI models rather than only buy.
Third-order effects
- If the pattern holds, local news consolidates around automation platforms supplying many titles at once, with quality control becoming the scarce editorial function — the failure mode Wired later documents at RuntimeWire, which favors quantity over quality shows what undisciplined versions look like.
- Disclosure norms split the market: some outlets label machine-generated work (as the LA Times does with AI ratings on opinion pieces), while others publish it unlabeled, pushing provenance verification onto readers and platforms.
The trend: News automation is expanding from templated data-driven stories to general-interest coverage at network scale, turning local publishing into a platform-supplied business where editorial labor migrates to oversight.