Likely AI, because otherwise they just made up some quotes on their own. Either is egregious journalistic malpractice. But we all know which one it was.
The maintainer reported in the comments of the article that he is exclusively misquoted in the 2nd half of the article (including misquoted calling himself a “gatekeeper”).
Edit: I did confirm all misquotes from archive versions of the blog and ars article. But I also read the blog yesterday and have memory and it hasn’t changed
Edit 2: Ars took down the article, replaced with archive.org. And here’s the blog archive version
Edit 3: they have posted a retraction notice and the original link now says the story is retracted instead of being a 404 https://arstechnica.com/staff/2026/02/editors-note-retraction-of-article-containing-fabricated-quotations/
Edit 4:
The maintainer, Scott, posted about getting AI zooped twice as well:
https://theshamblog.com/an-ai-agent-published-a-hit-piece-on-me-part-2/
Edit 5: the retraction notice is shit; doesn’t even say what was retracted. It also makes it sound like only the quotes were the only LLM-generated thing; it’s much much more likely the latter half of the article is LLM-generated in its entirety
Getting a “429 Too Many Requests” error.
But if true- this is depressing. I’ve always viewed Ars as a sort of standard bearer when it comes to tech news. Other outlets like The Verge do a lot of “best [product]” articles with affiliate links, and while I understand why, and don’t think they’re being dishonest, it just feels cheap. Seeing Ars fail at basic fact-checking is truly depressing.
They did not fail at fact-checking here. They straight up made shit up. Or rather, their AI chatbot did, but they published it, it got past their editors, it had the name of not just one but two of their writers on it.