The crass Fred Rogers of AI deepfakes has some Pittsburghers ‘horrified’

This story by Meg St-Esprit was originally published by PublicSource, a news partner of Technical.ly. PublicSource is a nonprofit media organization delivering local journalism for the Pittsburgh region at publicsource.org. You can sign up for its newsletters at publicsource.org/newsletters.

Lobbing curse-laden insults with TV’s famously serene painter Bob Ross.

Cracking jokes about school shootings. 

Being escorted in handcuffs by federal authorities. 

No, it couldn’t be Pittsburgh’s beloved icon Mister Rogers — the picture of moral clarity and togetherness. But it sure looks and sounds like him. What gives?

A growing wave of AI-generated deepfakes is distorting his likeness and warping a voice so many trusted.

Standing in contrast to these fictional scenes is the well-documented life’s work of the Latrobe native, whose educational advocacy reshaped the face of broadcast television and who died in 2003. Multiple organizations across the Pittsburgh region play vital roles in maintaining the legacy of Fred Rogers and his Neighborhood of Make-Believe. 

A pair of tennis shoes, a classic cardigan and original puppets are a few of the objects housed at the Children’s Museum of Pittsburgh, while the Fred Rogers Institute in Latrobe maintains more than 22,000 documents and artifacts. Fred Rogers Productions continues to create children’s programming inspired by the flagship TV show, and the Heinz History Center is trusted with preserving the WQED set. A prominent statue overlooking the meeting point of Pittsburgh’s three rivers stands as a testament to his lasting impact on the region.

Yet even as these community partners protect Mister Rogers’ story, a fast-moving threat is making that work harder. 

The videos were created using OpenAI’s new generative AI video app. Sora, launched in September, was downloaded 4 million times in the first month and is producing millions of fictional videos per day — including these videos of Fred Rogers doing and saying things inconsistent with his character. 

When Bethaney Tessitore first stumbled across one of these deepfakes during a mindless scrolling session, she was “horrified.” 

The Pittsburgh native, who now lives in State College, said “Mister Rogers’ Neighborhood” was integral to her childhood. “His biggest thing was kindness and just being kind to everybody, and it didn’t matter if people were different,” she said. Her two daughters, who joined her family at ages 8 and 9 via adoption, did not watch the show as young children, so Tessitore made sure to introduce her girls to the Neighborhood of Make-Believe herself. Over the years, they visited Idlewild, which has had several Fred Rogers-themed attractions, as well as the exhibit at the Heinz History Center. 

“That’s where my kids really got exposed to it, and I was teaching them a lot about him there,” Tessitore said. “It was just really interesting. I enjoyed it there; it kind of brought back memories of my childhood.” 

That’s why the Sora-generated video of Fred Rogers sharing lewd jokes with Tupac Shakur shocked Tessitore — it felt like an assault on those memories. She sought out a few more videos, and her feeling of unease grew. “I think I’ve seen my fill,” she said. “It kind of makes me mad that they’re trying to change his legacy. … He would be rolling over in his grave if he knew this.” 

‘Not what our future should be’

The institutions that are tasked with upholding Rogers’ life’s work are aware of the videos and concerned with what they convey. 

“Fred Rogers believed that there is a close relationship between truth and trust,” Fred Rogers Productions wrote in a statement to Public Source. “He earned the trust of America’s families through decades of dedicated study, scriptwriting, music composition and performance. When AI videos can be generated in seconds – making his likeness perform whatever tasks a user comes up with on a whim – it erodes the trust Fred worked a lifetime to create.” 

In a recent LA Times article, Fred Rogers Productions stated that they have contacted OpenAI and requested that his voice and likeness be blocked on the app. At the time of publication, videos of Rogers are still available on Sora, but no new videos involving his identity can be created. The videos have also migrated to other platforms including TikTok and Instagram, making them hard to moderate even if OpenAI changes its policies. 

Chelle Robinson, a producer on “Mister Rogers’ Neighborhood” from 1992 until its final season in 2001, said that anyone who knew Rogers in person would immediately recognize the videos as deepfakes, but worries it’s less apparent to the general viewer. 

“I was surprised at how well they had his mannerisms, honestly … That’s what makes it so hard, because it is convincing.” Robinson said she never even heard Rogers so much as swear, so the generated videos struck a chord. “I thought, ‘This is not what our future should be.’” 

Despite the concerns and legal questions about platforms’ use of celebrity identities, convincing AI deepfakes are here to stay. At the University of Pittsburgh, Associate Professor of Communication Jessica Ghilani teaches media studies. Ghilani said this technology is a kind of Pandora’s box. Once these videos are out there, they will influence the public perception of historical figures like Rogers. Even for viewers who know the clips are fake, viewing them has an impact. 

“It still diminishes our sense of trust in the visual realm,” Ghilani said, “and it can, in a worst-case scenario, lead to a broader sense of nihilism because if you don’t know what to believe, then it’s difficult to believe in anything.”

Curation vs. AI slop

Ghilani said there are few AI safeguards in place, and more are needed. “We don’t have regulations that are keeping up with the pace of the technology, and we have AI lobbyists that are actively arguing against it, saying that it will stifle innovation and profitability,” she said. “So I always want to point to regulation, but it’s difficult to feel like that is a realistic solution.” 

The legality of deepfakes is murky, as the technology races far ahead of the lawmaking process.

Just a few states have passed comprehensive laws on AI use. Most have at least seen the introduction of legislation.

Pennsylvania has not considered broad AI legislation, though the General Assembly  last year outlawed AI-generated child pornography and this year banned deepfakes or voice clones intended to defraud or harm people. The legislature is considering further efforts including a bill introduced in October to set rules for use of the technology in health care. 

On the federal level, the Trump administration is trying to bar states from regulating AI, arguing that a patchwork of rules would stifle innovation.

To the extent there is any legal consensus, it is that use of a deepfake is illegal when it is part of a crime like blackmail or fraud. If not, it’s generally legal. In the absence of a federal right to publicity, use of a person’s identity in a deepfake generally isn’t a copyright violation.

In the face of a lack of consistent regulation, individual corporations are left negotiating with companies like OpenAI.

Last week Disney announced an agreement allowing Sora to use animated characters, such as Mickey Mouse, Moana of the same-named film and Elsa of “Frozen”. Disney has placed limits around the use — disallowing videos that include drugs, sex or alcohol — in this landmark agreement in an emerging technological battlefield. Disney took a $1 billion equity stake in OpenAI as part of the arrangement.

Ghilani said viewers can take steps to personally counter the effects of misinformation online. She encourages lateral reading, or seeking out factual information from a trusted source, to combat the onslaught of negative images. She also suggests physical touchstones such as rewatching old episodes or visiting a Pittsburgh-area site committed to accurately preserving history. 

“Those material objects are so incredibly wonderful to get to see in person, and I do think that that can be a nice reality check for people,” Ghilani said. “If you feel like your brain has been fried by AI slop or by some kind of video about the person that you admire, then why not go to one of our wonderful museums and institutions and get to see some of that in person?” 

That’s the goal of curation, said Emily Ruby, the senior curator at the Heinz History Center who manages the Fred Rogers collection. 

“When I think about what I do as a curator, I’m collecting people’s stories.” Ruby conducts interviews, combs through original documents, and focuses on information that can be verified through trusted research methods. 

“I think it’s important to always go back to that source material and the person in the story to kind of try to counter all this vagueness out there,” Ruby said. “Remind people that museums and archives are holding these stories and actual records that you can trace back. You know it’s the truth; it’s documented.” 

Ruby said she’s been duped by some AI content on the internet herself, and encourages curiosity. “You just have to dig a little … It doesn’t usually take much. People just have to be more vigilant about what they are seeing.” 

A group effort to preserve Rogers’ neighborhood

As the History Center prepares to expand its physical footprint and create more space to display the entire WQED set, Ruby said she sees preserving Rogers’ legacy as a group effort. 

“As far as the people that are preserving his legacy, we’re kind of all in sync as far as how we want to share his story.” A large portion of visitors — like Tessitore — come due to nostalgia, and many are also introducing the younger generation to his work, as she did with her two daughters. 

Both types of visitors are crucial, Ruby said.

“The values that Fred represents and what he was trying to communicate in the show are something that’s universal and timeless.” She recalls how Rogers ended every episode by telling children they made each day special just by being themselves — instilling a sense of self-worth and value that has become synonymous with his memory. 

“I think it’s important that this message of his, that was really what was important to him, that we keep conveying that.” 

Even the AI videos themselves speak to his impact, Ruby said. 

“Why do people keep returning to him, quoting him? Why is he being reproduced in AI? Who cares about this television show personality 20 years later?” she said. “There’s a reason he’s being parodied. Because he is important.” 

This story was fact-checked by Rich Lord.