LANCASTER, PA – It’s hard to describe the city of Lancaster as anything other than quaint. Its vibrant but small downtown is peppered with coffee shops, bookstores and friendly locals. It’s got charm in abundance. The same can be said for the upscale Lancaster Country Day School, a K-12 private school in the area that hosts just over 600 kids. They frolic around in matching uniforms and play on crisp green grass once school lets out.

Perhaps that’s why the AI-generated sexual abuse scandal that recently rocked this town came as such a shock. No one saw it coming.

Only that’s not exactly true. Students did.

One student, in particular, was sent a pornographic deepfake of his upper school classmate on the communications app Discord, apparently in error. He deleted the photo, left the group chat, and filed an anonymous report to a state-run tip line, which in turn reported it to the school. But the school failed to act, according to lawyers representing at least 10 impacted families.

Over the next six months, two boys at the school continued to make AI-generated content of other young girls.

The deepfakes came to light in May 2024 when the school received additional information and filed a ChildLine report with the state, but a criminal investigation began only after parents notified law enforcement, according to an investigation by the Lancaster County District Attorney’s Office.

It was revealed that the two juvenile male students created and shared 347 AI-generated pornographic images and videos of 59 minors and one person over 18, according to the Lancaster County District Attorney’s Office. All of the victims were female; 48 were students at Lancaster Country Day School and the remaining 12 were acquaintances of the students who appeared in photos.

The two boys pleaded guilty to 59 felony counts of sexual abuse of children (manufacturing child sexual abuse material) and criminal conspiracy to commit that offense, Pennsylvania Attorney General Dave Sunday said in a press release. They’re set to be sentenced on Wednesday, March 25.

Yet the case at Lancaster Country Day School is hardly isolated.

Schools across the country are dealing with similar situations. As artificial intelligence platforms evolve, the prevalence of “nudify” applications, meant to remove someone’s clothing, is growing in tandem. School policies, legal recourse and awareness lag far behind, according to experts. That’s especially problematic for the often underage victims of this abuse – 90% of which are women – as studies show these deepfake nude images can have wide-ranging and long-lasting impacts.

“We continue to prioritize the health and well-being of our students,” Emile Kosoff, the head of Lancaster Country Day School said in a statement to USA TODAY. “Our deliberate and intentional approach aims always to ensure that our school community remains informed, continues to heal, and moves forward together.”

Some victims are ‘struggling just to get through’

For some of the young women at Lancaster Country Day School, the emotional impact has been profound. The creation and dissemination of the images occurred from October 2023 to May 2024, but much of the case is still playing out in court.

“A lot of them were sophisticated, college-bound, high-academic, functioning young women who also had all the right extracurricular activities to fill out all the right applications to go to all the right schools,” says Nadeem Bezar, an attorney and partner at Kline & Specter, who is representing at least 10 impacted families in a lawsuit against the school, which they intend to file after perpetrators’ sentencing. Those same students, Bezar says, “are now struggling just to get through.”

The families claim that the school did not take the appropriate steps to stop the creation of deepfakes after receiving the first tip. The school later dismissed its head of school and president of the board in the fallout.

Research shows the distribution of deepfake nonconsensual intimate imagery can lead to loss of privacy, psychological distress and reputational harm – and these symptoms can be experienced multiple times by an individual victim if that content is spread repeatedly.

“You have this future trauma, this future worry, ‘What if this shows up on my graduation day, what if it shows up when I start dating somebody else, what if it shows up on my first job interview, what if it shows up on my wedding day?’” explains Matthew Faranda-Diedrich, a partner at the Royer Cooper Cohen Braunfeld law firm, who previously represented some of the families from Lancaster Country Day School and is now representing another victim of deepfake abuse in the close-by New Hope-Solebury School District. “That is frankly one of the most damaging things because it’s really the thing that we have the least control over.”

The trauma can change how victims move through the world. For some of these young victims, images were lifted from their social media posts and transformed into pornography, which has exacerbated their emotional distress.

“You take images from totally innocuous, great situations that were positive memories for these girls – these minors – and now it’s turned into something nasty and gross,” Faranda-Diedrich says. “It totally ruins that memory of the nice moment with friends.”

Nina Jankowicz, an advocate for targets of online abuse who also had deepfake pornography generated of her, said there’s still a stigma surrounding those that speak out about this form of abuse.

“There’s a real disconnect. People don’t believe that the internet is real life,” Jankowicz says. “But the psychological effect of being depicted in digital sexual abuse like this changes how you move around in the world. Ultimately it does make you self-censor.”

That’s the “profound inequity in all of this,” Faranda-Diedrich says. His clients who experience deepfake abuse, primarily young women, are often made to “shoulder the burden.”

“It’s hard enough right now to be an adolescent, and then you add on top of it all these things, and it’s just so much for these young women and their families to bear,” he says.

Many schools are unprepared and the laws are still catching up

In 2024, a survey of 3,170 K-12 students, teachers and parents was conducted by the Center for Democracy & Technology to demonstrate the prevalence of AI deepfakes in schools and how prepared schools were to handle cases.

The survey found that 6 in 10 teachers were not aware of school policies and procedures for addressing authentic or deepfake sexual images, and only 16% said their school’s teacher training covered how to protect the privacy of a student depicted in a deepfake. Only 13% of students reported that their school has explained that sharing this type of AI-generated media is harmful to the person depicted.

Kristin Woelfel, coauthor of the survey, says schools shouldn’t have to experience an incident first-hand before taking preventative actions.

“We have numbers that we can put to this now and figures that do show how prominent this issue is, even if you’re not hearing these stories,” she says.

In 2025, the organization updated their data, finding that the percentage of students who had heard of deepfake and nonconsensual intimate imagery that depicts someone at their/their child’s school increased. However, an alarmingly low number of students know who to tell if they see or hear about these incidents (11%).

Woelfel says schools tend to have a greater focus on discipline than prevention – and also fall short on providing support to students after an incident.

“We did specifically ask parents what they want the students to have access to. In terms of providing resources to help students who were depicted, 77% of parents said that they wanted the school to provide this and only 8% of teachers were reporting that the school currently did that,” Woelfel said.

Lancaster Country Day School updated its reenrollment contracts to discourage students and families from publicly speaking poorly of the school, according to Bezar and a parent USA TODAY spoke with. The school did not return USA TODAY’s requests for comment about the contracts.

“The school knows that they have this deepfake issue, and they all of a sudden add this clause to their enrollment contracts,” Bezar says. “That to me seems a little disingenuous and unfair, and it doesn’t seem like someone’s apologizing.”

‘In wild times’: Deepfake incidents are growing

The number of people searching online for tools or platforms that can be used to create deepfakes is growing. And while the deepfakes themselves aren’t new, the scale at which they are being created is unprecedented, according to Charles DeBarber, takedown and digital victim advocate at DMCA Buddy.

“We’re kind of in wild times,” he says. Gone are the days of using CGI to stick a celebrity’s face on existing pornographic materials. “Now, you can do it with AI on really anybody pretty quickly.”

Between 2022 and 2023 Alice (formerly ActiveFence), a software development company, found that threads discussing and sharing “celebrity fake photos” climbed by 87%, and the number of threads related to the creation of sexual deepfakes depicting individuals rose by 400%.

Woefel says schools need to implement education on the harms and consequences of deepfake abuse to prevent future incidents. In her research, she found that none of these things were being conveyed to students.

“Or if they were, it was very few and far between,” she says.

Advocates and lawyers want to tighten mandatory reporting laws

Lancaster District Attorney Heather Adams determined that Lancaster Country Day School officials were not required to report the deepfake incident to the state-run tip line ChildLine or law enforcement. This is due to a limitation in the Pennsylvania Legislature that families and advocates are working to address.

Schools and teachers are mandated reporters who are required by law to report suspected cases of child abuse, neglect or maltreatment to authorities (such as child protective services or the police) when they have “reasonable cause” to suspect it.

However, there is a “loophole” in the mandated reporter statute that says you don’t have to report “child-on-child abuse,” according to Faranda-Diedrich.

“In this scenario … it’s not one of the things you could be criminally liable for,” he says.

As of Dec. 20, 2024, the Pennsylvania Legislature amended their laws to specifically include and define AI child pornography as child sexual abuse material (CSAM). Faranda-Diedrich is hopeful that amendments to the law regarding child-on-child abuse will be next.

Other states have similar limitations currently in place, and in several states, AI or computer-generated images are not included in existing child pornography statutes.

As victims struggle to remove content, families want a focus on prevention

In May 2025, the Take It Down Act was signed into law to combat nonconsensual intimate imagery, including deepfakes and revenge porn. The law requires social media platforms and similar websites to remove nonconsensual intimate imagery, including pornographic deepfakes, within 48 hours of a verified request from a victim.

However, victims still struggle to get their photos taken down, increasing the likelihood that images will continue to spread and retraumatize them. A 2025 report from the Center for Democracy and Technology found that across eight popular social media and online platforms, the policy language and structure for reporting the nonconsensual distribution of intimate images was vague and inconsistent. Platforms provide “limited transparency and support,” the report says, and “fail to provide real-time tracking of reports [and] reporting outcomes,” making it more difficult for victims to navigate their case.

The generational tech gap is also “setting victims up for failure,” says victim advocate DeBarber. One of his college-aged clients had to meet with the dean of her university and explain AI-generation to them, using pictures of cats to demonstrate how images can be altered.

“They’re expected to be their own leading advocate most of the time,” he says. “Schools need to get these policies in place, so especially some of the worst perpetrators can’t go, ‘Well, I didn’t know this was wrong, I didn’t know this was bad.’”

That’s the thing parents at Lancaster Country Day School wish the most – that prevention had been in place – or at least that action would have been taken after the first image was discovered.

“The school should be instilling in these children what is right and wrong with using AI,” Faranda-Diedrich says. “They have to broaden the lens and include the perils of AI for adolescents, just like they talk about perils of drug use or of promiscuity. It’s a peril of being a young person.”

This story was supported by a grant from the Tarbell Center for AI Journalism. Funders do not provide editorial input. 

This article originally appeared on USA TODAY: The deepfake AI porn scandal that shocked a small town, private school