Build-A-Bear gets a satanic AI makeover
PLUS: No subsidy for fast-delivering French mothers; a college website gets taken over by AI
Welcome to NewsGuard's Reality Check, a report on how misinformation online is undermining trust — and who’s behind it.
Today:
Grizzly images: AI-generated “satanic Build-A-Bears” fool TikTok users
Birthing a hoax: Paris maternity myth debunked
College news website hijacked to become an AI-generated content farm
And more…
Today’s newsletter was edited by Jack Brewster and Eric Effron.
1. The Bear Truth: TikTok Users Ensnared by Satanic Build-A-Bear Spoof
By Coalter Palmer and Jack Brewster
As if bears weren’t scary enough…
What happened: TikTok users are citing AI-generated images, originally shared by a popular Facebook satire page, as evidence that popular stuffed animal chain Build-A-Bear has undergone a “satanic” rebrand and is offering a new line of “Baphomet Bears.” (It’s not clear which AI image generator was used to create the fake photos but they seemed so convincing that we decided to try it ourselves; see below.)
Baphomet is a pagan deity that the Knights Templar were accused of worshiping. It later became an occult symbol used by groups including the Church of Satan, according to Encyclopedia Britannica (Trust Score: 87.5/100).
A closer look: The AI-generated images, first fact-checked by USA Today (Trust Score: 100/100), originally come from a March 27, 2024, post by the self-described satire Facebook page “The Pumpkin Express,” which boasts 123,000 followers.
The Pumpkin Express captioned the photos: “BUILD A BEAR for you or the kids!!! These stores are getting with the times! I want to build a baphomet baby bear. He shall be named Baphy. … #relaxitssatire.”
Some users didn’t get the joke…
Ten days later, on April 6, 2024, the images went viral on TikTok when user @officialtinye5 shared The Pumpkin Express’s images in a post that included no indication that the images were originally intended as satire. Watch the video below…
Another TikTok video posted by creator Ihsaan Abdul Maani, a user with 154,000 followers, then captioned a slideshow of the AI images, “Build A Bear now has a Baphomet Bear #buildabear #endtimes.” Watch the video below…
Actually: True Media, an AI image detection software, identified all six images as having been AI-generated.
A Build-A-Bear spokesperson told NewsGuard in an April 2024 email that the chain “currently does not, has not, nor has any plans to offer the products or services in question.”
Some context: This isn't the first time that AI-generated photos are being used to suggest that major brands are promoting satanism. In the past year, NewsGuard has documented other instances where social media users created AI images to support claims that:
McDonald’s is selling a Satan-inspired Happy Meal
Arts and crafts store Hobby Lobby sells Satanic statues
Target sells Satan-themed clothing to children
See below…
We decided to test it ourselves. We instructed OpenAI's DALL·E 3, an AI image generator, to create an image depicting a "satanic-themed NewsGuard newsroom," complete with "Baphomet near the door." Below is the image it produced:
Click here to find out more about NewsGuard Trust Scores and our process for rating websites. You can download NewsGuard’s browser extension, which displays NewsGuard Trust Score icons next to links on search engines, social media feeds, and other platforms by clicking here.
2. False Start: No Parisian Policy for Pre-Olympic Births
By Macrina Wang
French mothers-to-be may proceed with their pregnancies as planned.
What happened: The baby boom that wasn’t. Accounts on TikTok, X, and Facebook have falsely claimed that the French government is incentivizing pregnant Parisians to give birth before the Paris Olympics begin.
The supposed incentive is to free up hospital beds ahead of the games.
A closer look: As evidence, these accounts cited a fake TF1 broadcast (TF1 is the largest TV network in France).
Watch the fake broadcast below:
Translation of the on-screen text: “Parisians are invited to give birth before the start of the Paris Olympic Games. The Ministry of Health has issued a recommendation for Parisian women who plan to give birth at the end of July 2024. According to the recommendation, women whose due date is between July 22 and 28 are expected to give birth before July 26. If the delivery takes place before or on July 25, the state will reimburse up to 50% of medical costs to women.”
The text continued, “For women who are ready to transfer childbirth to other regions of France for a specified period, compensation of 25% for medical fees is provided.”
Actually: A NewsGuard search of France’s Ministry of Labor, Health and Solidarity website found that there were no announcements encouraging Parisians or offering incentives to give birth prior to the Olympics.
France’s Ministry of Labor, Health and Solidarity told Agence France-Presse (AFP), which previously fact-checked the claim, that everything said in the video "is completely false.”
TF1 also told AFP that the video was not theirs.
Who’s behind it: It is unclear where the fabricated TF1 video originated. The earliest instance of the video identified by NewsGuard is from an April 9, 2024, Telegram post by pro-Kremlin account @BorisKarpovRussie, who claims in his bio that he is “censored in France.”
The post said, “Western values: ‘Give birth faster!’ .... The Olympics has made its own adjustments to French women's due dates.”
On the same day, Pravda-fr.com (Trust Score: 7.5/100), a website that is part of a Moscow-based news network targeting Ukraine and Western Europe, posted an article that repeated @BorisKarpovRussie’s Telegram post word-for-word.
Do you work in Trust and Safety for a technology company, in brand safety for advertising or otherwise counter misinformation as part of your job? Find out about NewsGuard’s weekly Risk Briefings, a more detailed briefing for professionals. Click here.
3. And One More Thing: School Newspaper Taken Over by AI Content Farm
What happened: After the University of Michigan-Flint decided to close its student newspaper and allowed the domain name The MichiganTimes.com to expire last year, an AI content farm snatched up the domain and began operating a new website under the guise of a newspaper affiliated with the school.
A closer look: As the Flint-based East Village Magazine reports, while the new site — operating under the same “The Michigan Times” name — “resembles a slickly produced version of the old student newspaper site,” university officials have confirmed it has “no connection to UM-Flint” and “also appears to be misrepresenting its ties to alumni.”
“What it does have,” East Village Magazine writes, “are many of the characteristics of an automated site that pulls stories from legitimate news sources; quickly scrambles and rewrites them using artificial intelligence (AI) chatbots; and publishes them under a different byline.”
You can read East Village Magazine’s story, which cites NewsGuard’s extensive reporting on AI content farms, by clicking here. These sites use AI chatbot prompts to generate or rewrite news stories, sometimes introducing false claims and made-up citations, cheaply to produce their own content. Many then generate revenues by signing up to have advertisements placed on these sites automatically through algorithms, powered by companies like Google that provide these programmatic advertising services.
It’s another reminder about the threat posed by AI content farms, which are rapidly spreading across the internet, and, as NewsGuard showed in recent commentary for The Wall Street Journal, cheap and easy to launch on your own.
Produced by co-CEOs Steven Brill and Gordon Crovitz, and the NewsGuard team.
We launched Reality Check after seeing how much interest there is in our work beyond the business and tech communities that we serve. Subscribe to this newsletter to support our apolitical mission to counter misinformation for readers, brands, and democracies. Have feedback? Send us an email: realitycheck@newsguardtech.com.