Several news organizations have deleted articles written by suspected freelance journalists that are now believed to have been generated by AI.
On Thursday, Press Gazette reported that at least six publications, including Wired and Business Insider, have deleted articles from their website in recent months after it was discovered that a story written under the name Margaux Blanchard was generated by AI.
Wired released a story in May titled “I fell in love playing Minecraft. The game became a wedding venue.” A few weeks later, the outlet defeated the story with an editor's note. “After further reviewing the article, the wired editorial instructor determined that this article did not meet the editorial standards.”
The story cited “Jessica Hu” according to Wired Gazette.
Both the press's official gazette and the Guardian were unable to verify Hu's identity.
Press Gazette reported that in April he published two Blanchard essays and two Blanchard essays.
Earlier this week, Business Insider deleted the article after Press Gazette warned the outlet about the author's reliability. Both article pages contain a message saying “Because the story has been deleted.” [they] We didn't meet Business Insider standards.”
The Guardian reached out to both Wired and Business Insider for comment.
Press Gazette says he was first warned of fraudulent conduct in Blanchard's article by Jacob Ferdi, editor of a new magazine called the Dispatch. Furedi said earlier this month that he received a pitch from Blanchard saying “Gravemont, a decommissioned mining town in the countryside of Colorado, has been reused as one of the world's most secret training grounds for death investigations.”
On the pitch shared with the Press Gazette, Blanchard wrote: “I want to tell the story of a scientist, a former Coppus, and a former miner who now deals with the dead every day. I explore the ethical ti-tropes that use real humans in staged environments, not as mourning, but as true archivists.
“I'm the right person for this because I have previously reported on hidden training sites, have clearance contacts in forensic circles, and know how to navigate sensitive, closed communities with empathy and discretion,” she added.
Furedi told the spokesman that Pitch sounded like it was generated by ChatGpt and could not find any details about Gravemont. The Guardians were also unable to confirm details of the suspicious town.
When I asked Blanchard how she discovered the town, she replied. “I'm not surprised you couldn't find much. Greigmont doesn't promote himself. I first interviewed a piece that was irrelevant to a retired forensic pathologist.”
She went on to say: “Over the next few months, I stitched it together further by requesting public records, talking with former trainees, and tips buried in forensic association meeting materials, none of which were mentioned by the name of the print.
“This is one of the places that exist in the collective memory of the industry, but it's under enough radar to avoid coverage, so I think it resonates with the dispatched readers,” added Blanchard.
Furedi told Press Gazette that she knew she was “bummed” despite the fact that the pitch was “very convincing.” He asked Blanchard to see her standard rate and how long she will spend on the field.
In response, Blanchard ignored Furedi's request to see the request for public records, saying instead he would “ideally spend five to seven days on the ground” and would be paid around $670.
Last Friday, Furedi denounced Blanchard in an email saying she would publish a false story she had not responded to. Press Gazette further reports that Blanchard did not respond to his own request to provide evidence that she is a real person.
This case of false AI-generated reports follows the May error when Chicago Sun-Times Sunday paper ran a syndicate section containing fake reading lists created by the AI.
Marco Buscaglia, a journalist who worked for King Feature Syndicate, turned to AI to generate the list, saying: [an AI program] Spit out… Usually that's something I don't… Even if I'm not writing anything, at least I'm sourced and vet it right and making sure it's all justified. And I definitely failed that task. ”
Meanwhile, in June, the Utah Court of Appeals approved the attorney after it was discovered that he used ChatGpt to refer to a nonexistent trial.
