Search Google News and you'll find AI-generated articles brazenly stolen from legitimate media outlets. The Post has already identified at least one such plagiarism of its published work.
Late last week, the Post tested Google News, searching for recent articles about Federal Trade Commission candidate Melissa Holyoke and sorting the results by most recent publication date.
The Post's Jan. 8 exclusive article on Holyoak was listed lower in search results than a near-identical copy — with the common name “Business News” and the bizarre domain address “biz.crast.” The number of AI-generated articles published by “.net'' news organizations that appear to be mass gathering large amounts of information.
The fake version of the article featured the same artwork, and the regurgitated copy even included a reference to “The Post.” The plagiarism was attributed to “Sean Johnson,” whose byline has more than 17,800 pages of results and dozens of articles published last Friday alone.
By late Friday afternoon, a Google spokesperson confirmed that the article “violates our policies and will be removed.”
Google also said that while AI-generated content does not violate its policies, content may be removed if it is determined to be “spam” published specifically with the intention of ranking high in news results. It was acknowledged that there is a possibility that
Independent Outlet 404 Media Last week, he introduced a screen in which AI-generated plagiarism, including a Star Wars post published by Distractify and a regurgitated article about execution-style murder published by Heavy.com, were displayed alongside the real thing. I raised this issue after getting the shot. The article will appear in Google News search results.
According to News/Media Alliance CEO Danielle Coffey, the proliferation of AI-generated articles is already a “real problem” for the industry.
“This is a broken system because quality human-generated content is not being rewarded across the board,” Coffey told the Post. “It would create a situation where it would be impossible to generate the revenue to keep newsrooms afloat.”
Coffey was one of several experts testifying before a Senate committee earlier this month about the dangers AI could pose to the future of journalism. At the same event, Condé Nast CEO Roger Lynch said AI chatbots are “made with stolen goods” and should be regulated by Congress.
Digital Content Next CEO Jason Kindt says AI-generated news content is proliferating alongside real articles on platforms run by Google, which controls 90% of the online search market. It's very worrying.”
“Google has incredible market power in discovering news brands and making them available to the American public,” Kindt said. “AI is a tool and perhaps should not be seen as the bad guy here. Rather, it is an accelerant of risk if the power of gatekeepers is left unchecked.”
As The Post has reported, media companies have been using AI-powered chatbots, such as OpenAI's ChatGPT, to lift copyrighted works in recent months without proper credit or compensation. expressing anger towards.
Last month, the New York Times filed a major copyright infringement lawsuit to protect its business model, while other newspapers were embroiled in intense negotiations to secure payment for their content through licensing deals. It is.
404 Media's report cites several specific examples of blatant plagiarism, in which Google posts that feature headlines, photos, and text that are identical or nearly identical to authentic media articles, seemingly using AI He said that he was “advertising” the information by redirecting it to a site that utilizes it.
The Google News example was discovered by searching for related topics and setting search parameters to content published in the past 24 hours, the report said.
“The presence of AI-generated content on Google News points to two things: first, the black box nature of Google News, and entry into Google News rankings is opaque to begin with; But it's clearly a gameable system,” said Joseph Cox of 404 Media. I have written.
“Second, in an age of consumer-accessible AI, Google may not be ready to moderate its news service, which essentially allows anyone to use AI with little regard for quality or originality. “We'll be able to churn out a lot of content without having to worry about it,” Cox added. .
Danny Sullivan, head of Google Search, questioned the report's methodology. In a long X thread, I argued that sorting search results by date is “explicitly asking the system to ignore normal relevance rankings.”
A Google spokesperson also denied the report's findings in a statement to the Post.
“Claims that these sites were featured prominently in Google News are not accurate. The sites in question were only featured for artificially narrowed queries, such as queries that explicitly excluded the date of the original article. displayed,” a spokesperson said in a statement.
A Google spokesperson said: “We take the quality of our results very seriously and have clear policies against content created with the primary purpose of ranking highly in news. We are removing sites that violate the rules.”
Google also argued that it was unfair to claim that an article was “boosted” just because it appeared as a result of responding to a specific user's request.
Asked for comment, 404 Media's Cox responded that his article was “clear” and “contains no factual errors.”
Jason Kabler, co-founder of 404 Media, also defended the report in a statement, saying, “Normal search terms with no time limit will still appear above sites that plagiarize this content in Google News.” he claimed.
