"They hate the news. They view the news as a necessary evil within a wider gambit to deploy adware, malware, pop-ups, and auto-play video."
That is Cory Doctorow's assessment of why media executives are rushing to replace journalists with AI, and it is hard to argue with the evidence he lays out.
The Summer Reading List That Didn't Exist
The most damning example: Hearst distributed a syndicated summer reading guide across its newspaper network featuring book titles that do not exist. The piece was credited to freelancer Marco Buscaglia, who had been given the job of producing what would normally require dozens of writers, editors, and fact-checkers. One person. A 64-page insert. As reported by 404 Media's Jason Koebler, this setup was designed to fail.
Nobody at Hearst caught the hallucinated books because nobody at Hearst was checking. The AI did not cause the quality collapse. It just removed the last friction that had been preventing it.
The Pattern Is Older Than ChatGPT
Doctorow draws a straight line from AI journalism to the broader history of corporate cost-cutting in media. Customer service got outsourced to call centers, then replaced with chatbots. Not because AI handles problems better, but because the companies had already decided that solving customer problems properly was not worth paying for. The same logic applies to newsrooms.
Media companies have spent two decades hollowing out editorial staff, replacing beat reporters with content farms, and plastering every article page with so many ads and tracking scripts that the actual text is an afterthought. AI-generated articles are not a technological innovation in this context. They are the logical endpoint of treating journalism as a cost center rather than the product.
For anyone who works with AI tools daily, this is a useful distinction to keep in mind. The technology itself is neutral. A skilled writer using Claude or ChatGPT to research faster, draft outlines, or check facts is doing fundamentally different work than a media conglomerate using AI to eliminate the writers entirely. The first is augmentation. The second is abandonment dressed up as automation.
The problem was never the AI. The problem is that the people making deployment decisions do not value what they are automating away.