Wired: Editorial offices secretly deploy AI assistants for article writing — and remain silent about the consequences

AI writing quietly penetrates newsrooms—not through grand announcements, but through small concessions: first a news brief draft, then a financial data summary, then an initial version of a full report. Publishers call it efficiency. Journalists see something different. Wired raises a topic the media industry prefers to avoid: what happens to journalism when AI starts taking the first—and most crucial—step in creating content? A draft shapes structure. A draft determines angle. A draft is already half the finished article. If a machine writes it, who bears responsibility for what appears in print?

Pressure on newsrooms is mounting from several directions simultaneously. Ad revenue continues to decline. Readers increasingly get information from social media and search engines—often bypassing publisher sites entirely. Against this backdrop, reducing text production costs looks attractive: AI tools promise the same speed with lower personnel expenses. BuzzFeed, CNET, Sports Illustrated, G/O Media—all have experimented with text automation over the past two or three years. Results are mixed: there have been factual errors, outright failures, and quiet withdrawals from experiments without official explanation. But the problem goes deeper than factual errors in AI drafts.

Journalism exists not because you need to quickly gather facts into a paragraph. It exists because a living person with experience, sources, and a perspective can find a story where no one is looking, ask an uncomfortable question, and take responsibility for every word. AI can imitate text structure.

It cannot imitate judgment. There's another hidden cost—reputational. When it becomes clear that a piece was written or assembled by a machine, reader trust in the publication drops—and recovers extremely slowly.

CNET faced this in 2023, when dozens of financial pieces were found to have been created by AI without proper disclosure. The editorial team was forced to retract some publications and introduce a mandatory label. The lesson proved painful and expensive.

Technology companies selling tools to newsrooms prefer different language. They speak of "accelerating workflows," "reducing cognitive load," "assisting, not replacing." In practice, the boundary between "assistance" and "replacement" is blurred and constantly shifting. Today AI writes a news feed draft. Tomorrow—an analytical review. The day after—an opinion column with a human author's name on the byline.

The journalists' own reaction is telling. In newsrooms where AI tools are implemented without open discussion with the team, hidden resistance grows. People feel their professional identity is being eroded. When your job is to edit someone else's draft rather than create your own, you stop being a journalist in the full sense.

What is happening is not simply a matter of technology or economics. It is a question of what journalism is as an institution. If newsrooms embrace AI writing as a norm under pressure from short-term financial incentives, they risk destroying what distinguishes them from algorithmic news aggregators. Readers choose living publications not for speed—algorithms won. They choose them for trust, for voice, for the judgment of a specific person. This cannot be automated—and attempts to do so invisibly sooner or later become apparent.