Popular Tech Site Has Been Publishing AI-Written Articles For Months
By Alexa Heah, 12 Jan 2023
While New York City’s Department of Education recently banned the use of artificial intelligence (AI) chatbot ChatGPT, it seems news publishers could take a cue from the authorities, as Futurism reveals a popular technology site has been posting articles that weren’t written by humans.
CNET, a go-to news outlet for those looking for the latest updates in technology, has been said to be making use of “automation technology” to write its articles for months, allegedly beginning in November 2022.
The AI-written articles were first spotted in a tweet by online marketer Gael Breton, who noted that a post on the page had a disclaimer that mentioned it had been written by a bot, but said the information was “thoroughly edited and fact-checked” by an editor.
Most of the articles written in such a manner were filed under the authorship of ‘CNET Money Staff’, and revolved around financial topics such as ‘What Is A Credit Card Charge-Off?’ and other explainers.
While there isn’t a rule against AI-automated publications, it does seem that CNET had tried to keep its use of the technology on a down-low, not offering an official announcement on the matter despite having published 73 such articles.
Will this erode the trust of readers on news sites, or will it be hard for the layman viewer to spot an AI-written article from one penned by a human? This certainly isn’t the first time news outlets have used AI to write articles, but it could lead to more such practices in the industry.
It looks like the experiment started around November 11 based on publish dates.— Gael Breton (@GaelBreton) January 11, 2023
For now, it's a small-scale test as we only see 72 results disclosed as written by AI. pic.twitter.com/KuqSCQEpUp
As Futurism notes, famed news agency The Associated Press, which claimed to be one of the first news organizations to dabble in AI, has been using bots to fill in its thousands of earning reports since all the way back in 2015.
The difference is that the articles on CNET aren’t just predetermined layouts being filled in by bots, but rather, contain a wider range of vocabulary and sentence structures that point towards the use of a program like ChatGPT.
In addition, there comes the issue of search engine rankings for AI-written content, which may be deemed as “low effort” compared to original, human articles. While Google promised to penalize AI-generated content, Breton’s searches seemed to point otherwise.
And it looks like Google is not giving a shit about it, rewarding several of these pages with great search traffic.— Gael Breton (@GaelBreton) January 11, 2023
Soooo.. Is AI content ok now? ð¤
I'd love to hear your thoughts below. pic.twitter.com/GgsqU9eNIH
As for CNET, its website states that it will continue to publish articles with “editorial integrity,” and that its editorial guidelines remain guided by “accuracy, independence, and authority.” Will this set a precedent for other news outlets going forward? Perhaps it’s only inevitable.
More related news