This article was AI-generated as part of an experimental historical-content project. The date reflects the period being analyzed rather than the date the article was originally written.
There’s an assumption built into a lot of PR planning that coverage in a large, established publication is a durable asset. Get the story placed in a recognizable outlet and it keeps working for years: in search results, in sales decks and, eventually, as a citation on Wikipedia.
This week offered a useful correction. Wikipedia’s editors have downgraded CNET.
As Ars Technica reported on Thursday, following a Futurism investigation, Wikipedia’s Reliable sources/Perennial sources list now splits CNET into three periods. Before October 2020, it is considered generally reliable. From October 2020 through October 2022, after Red Ventures acquired the site, editors note a “deterioration in editorial standards” and record no consensus on reliability. From November 2022 onward, CNET is listed as generally unreliable, after it began using an AI tool “to rapidly generate articles riddled with factual inaccuracies and affiliate links.”
CNET told Futurism it is not actively using AI to create new content. The rating changed anyway.
The myth of permanent authority
The Perennial sources page is one of the less glamorous parts of Wikipedia, and one of the more consequential. It summarizes the community’s past discussions about whether a publication can support claims in articles. It isn’t a formal ranking, and editors still weigh context case by case. In practice, though, a “generally unreliable” label means a citation to that outlet is likely to be questioned, replaced or removed.
What stands out about the CNET entry is that it’s date-bound. Wikipedia didn’t decide CNET was always bad. Editors decided the same brand produced different quality at different times, and recorded where the lines fall. A CNET article from 2018 and one from 2023 now carry different weight, even though they share a domain and a logo.
That’s a more realistic model of source reputation than most media lists use. PR teams tend to treat outlets as fixed tiers. Wikipedia’s editors treat them as institutions whose behavior can change, and whose reliability gets revised when it does.
How AI content got here
The trigger is specific. CNET’s AI-written finance explainers, published under a “CNET Money Staff” byline, drew attention in January 2023 for errors and plagiarism. Wikipedia editors began discussing the outlet’s reliability shortly afterward. According to Ars, the debate extended to other Red Ventures properties, such as Bankrate and CreditCards.com, and editors criticized the company for not being forthcoming about where and how AI was used.
That last point deserves attention. The cost didn’t come only from the errors. It came from uncertainty about which articles were affected. When readers can’t tell where a problem starts and stops, a cautious community draws the boundary wide.
Why this matters beyond publishers
For a company or an executive, the practical issue is that earned media is only as useful as the standing of the outlet that published it, and that standing is openly negotiated. A profile in a downgraded source may still rank on Google and still look impressive in a press kit. On Wikipedia, it may no longer hold up a sentence.
And Wikipedia’s sentences travel. They inform Google’s knowledge panels and sit in the material AI systems learn from and retrieve. As AI tools cite Wikipedia directly, the community’s sourcing judgments shape what those systems repeat. Editors are also still debating their own rules for AI-generated text, which suggests source disputes like this one will keep coming.
So when third-party coverage matters, it’s worth checking how Wikipedia’s editors currently regard the outlet, whether it has recently changed ownership or content strategy, and whether your most important outside validation is concentrated in one place.
Reliability ratings are slow, public and argued over on talk pages. That is exactly why they work as infrastructure.