Seven people quoted in a newsletter about journalism practices say they never said the things attributed to them. The quotes were generated by ChatGPT and Perplexity, and the author - a veteran media executive - published them without checking.
Mediahaus, the European media group that publishes the Irish Independent and Sunday Independent, has suspended Peter Vandermeersch from his "Journalism and Society" fellowship. Vandermeersch previously served as Mediahaus's Ireland chief executive from 2022 to 2025 and before that was editor-in-chief of NRC, a Dutch newspaper also owned by Mediahaus.
The fabricated quotes were discovered after a staff member at NRC flagged that statements in Vandermeersch's newsletter couldn't be verified. All seven quoted individuals subsequently confirmed they never made the remarks attributed to them.
How a Newsroom Veteran Fell for Hallucinations
Vandermeersch admitted he relied on summaries from large language models, "trusting they were accurate." He described falling into "the trap of hallucinations" - the well-documented tendency of AI chatbots to generate plausible-sounding but completely invented information, including fake quotes with real people's names attached.
This is notable not because a journalist used AI (many do, and openly), but because of who got caught. Vandermeersch isn't a junior reporter cutting corners on deadline. He's a former editor-in-chief and C-suite executive at a major European media group, serving in a role specifically focused on journalism standards.
Mediahaus CEO Gert Ysebaert said the situation violated company policy: "At Mediahuis, we apply strict rules for the use of AI, where diligence, human oversight and transparency are essential."
The Real Lesson for Anyone Using AI for Research
The failure pattern here is one every AI tool user should recognize. LLMs are confident-sounding research assistants that will invent sources, quotes, and citations with the same authoritative tone they use for accurate information. There is no visual difference between a real quote and a hallucinated one in ChatGPT's output.
The practical rule hasn't changed: AI tools are useful for drafting, summarizing, and brainstorming, but any factual claim - especially direct quotes from named individuals - needs verification against primary sources. That applies whether you're writing a newsletter, a blog post, or a client report. The tools don't know when they're making things up, and neither will you unless you check.