■ CRITICAL RISK ■ Media & Communication
AI won't replace journalism, but it's gutting journalists — commodity news writing is automating fast in an industry that was already broke. Original reporting, source cultivation, and accountability work survive; the jobs that were mostly rewriting press releases do not.
“AI writes breaking news before you finish your morning coffee.”
Our AI replacement risk score — how we score jobs
Be honest about what much of the job became before AI showed up: aggregating wire copy, rewriting press releases, churning SEO posts about what a celebrity said on a podcast, and summarizing earnings reports. Newsrooms automated earnings and sports recaps years ago with template systems; language models now draft anything derivative — and derivative content is, uncomfortably, a large share of published 'journalism.' AI also compresses the honest grunt work: transcription that once ate afternoons takes minutes, document sets get summarized instantly, and data analysis that required a specialist is increasingly conversational.
What AI cannot do is the part that defines the trade: getting a source to tell you the thing they're not supposed to tell you. Working a beat until officials return your calls, filing records requests and knowing what to ask for, standing in a courtroom or a flood zone, verifying that the leaked memo is real, and taking legal responsibility for publishing it — none of that is a text-generation problem. Investigative and accountability journalism may actually get more powerful, with AI as a research assistant that never sleeps.
The wound is economic as much as technological. The industry's business model was collapsing before AI — decades of layoffs, dead local papers, platform-dependent traffic — and now AI answers in search results siphon the clicks that funded what remained, while AI-generated content floods the same channels. Our 78 is a bet that employment keeps falling even though the craft endures: fewer, better reporters doing original work, atop a shrinking pile of money, while the aggregation layer beneath them gets fully mechanized.
Automatability: our editorial assessment of current and near-term AI capability
Already deep into it — newsroom contraction predates AI by two decades, and AI is accelerating the cuts while search-traffic erosion starves the business model further. Commodity-content roles are disappearing right now. Reporters with beats, sources, and investigative chops will remain employable through the 2030s, but at far fewer outlets, with nonprofit and subscription models carrying more of the weight.
No — but it doesn't need to in order to devastate employment. AI already produces the aggregation and rewrite content that filled much of the industry's page count, and it's eroding the search traffic that funded newsrooms. What survives is original reporting: sources, verification, investigation, presence. The profession contracts toward a smaller number of people doing genuinely irreplaceable work.
Only with eyes open. The economics are brutal and getting worse, and entry-level aggregation jobs — the traditional first rung — are precisely what AI eliminates. If you go in, go in to report: build a beat, learn records requests and data skills, and develop a direct audience early. Journalism as a craft has a future; journalism as a stable mass employer does not.
Quite a lot of the routine layer: drafting recaps from structured data, summarizing reports and press releases, transcribing interviews, translating, generating headlines and social copy, and first-pass copy editing. Many newsrooms use it for research assistance and document analysis on large leaks. What it doesn't do is originate: no source relationships, no eyewitness presence, no accountability for getting it wrong.
Own things AI can't: a beat where officials know your name, a source network, subject expertise deep enough to catch what generic coverage misses, and ideally an audience that follows you rather than your outlet. Use AI tools to multiply your reporting capacity instead of competing with them on volume. The reporters most at risk are the ones whose bylines could be swapped without anyone noticing.