■ HIGH RISK ■ Media & Communication
AI can already do the copy-editing, headline-writing, and layout that filled an editor's day — and the industry's economics were shrinking the desk before AI arrived. Editorial judgment survives; the number of chairs it occupies doesn't.
“AI curates and edits news. Your red pen is a relic of journalism's romantic past.”
Our AI replacement risk score — how we score jobs
Newspaper editors do far more than fix commas: assigning and shaping stories, line-editing for clarity and legal risk, writing headlines that are accurate and clickable, managing reporters, deciding what leads, and serving as the last line of defense against libel and error. But the daily mechanics — copy-editing, headline variants, SEO optimization, layout, wire-story selection, and pushing content across print, web, and social — consume most desk hours.
Those mechanics are exactly where AI has landed. Language models copy-edit competently, generate headline options tuned for search, summarize wire copy, and repackage a story for every platform in seconds; newsroom CMS vendors are shipping these features as standard. Wire services and major outlets have used automated writing for earnings and sports briefs for years, and curation — deciding what surfaces where — increasingly runs on engagement algorithms rather than a news-meeting argument. Layer this on an industry that has been cutting desk jobs for two decades, and the copy desk, already gutted by consolidation, is the obvious next efficiency. Our 72 reflects AI accelerating a contraction that was well underway.
What resists automation is accountability and judgment. Someone must decide a story is fair, a source is solid, a headline won't get the paper sued, and a tip is worth three weeks of a reporter's time — and that someone must be able to take the blame. AI-generated errors in published copy are a legal and reputational nightmare, which is why even AI-enthusiastic outlets keep humans on final approval. Investigative editing, mentoring reporters, and community trust don't compress into a feature. The future is fewer editors doing higher-judgment work atop heavily automated production — a real job, held by far fewer people.
Automatability: our editorial assessment of current and near-term AI capability
The pressure is current, not projected. Newsroom AI tools handle copy-editing, headlines, and repackaging today, and an industry that never stopped cutting will bank those savings immediately — expect desk roles to keep thinning sharply through the late 2020s. Senior editors with final-approval authority and investigative oversight remain, but the traditional ladder of copy-desk jobs beneath them is being sawed off now.
As a lifelong copy-desk career, honestly, no — the industry was shedding those jobs before AI and is shedding them faster now. As an editorial-judgment career, yes, but the path is narrower: fewer positions, concentrated at outlets with resources, weighted toward standards, investigations, and management. Anyone entering should build skills that transfer beyond newspapers.
It copy-edits well, writes serviceable headlines, and summarizes accurately most of the time — 'most of the time' being the problem. It also misses factual errors it has no way to detect, can't judge fairness or source reliability, and introduces confident mistakes that would be catastrophic in print. That's why the emerging model is AI doing production passes with a human owning final approval.
Attach yourself to the parts AI can't absorb: legal and ethical review, investigative story development, staff management, and community relationships. Learn the AI tools well enough to supervise them — editors who can design and audit automated workflows become more valuable, not less. And keep a foot outside newspapers, because the durable editorial jobs are increasingly elsewhere.
Fully automated news exists today for commodity content — earnings, scores, weather — and it's expanding. But original reporting requires sources, access, and legal accountability no model provides, and outlets that publish unreviewed AI copy keep generating public corrections that prove the point. Newsrooms shrink and restructure around AI; the reporting-and-judgment core persists, smaller.