Teens get probation after using AI to create fake nudes of classmates

LANCASTER, Pa. (AP) — Two teenage boys who used artificial intelligence to create fake nude photos of their classmates at an exclusive private school in Pennsylvania received probation Wednesday after dozens of victims described the images’ traumatizing effect on them.

The boys were 14 at the time. They admitted this month that they made about 350 images, showing at least 59 girls under 18, along with other victims who so far have not been identified.

Authorities and the girls themselves said the boys took images of the girls from school photos, yearbooks, Instagram, TikTok and FaceTime chats in 2023 and 2024, and morphed them with images of adults depicting nudity or sexual activity.

More than 100 students and parents from Lancaster Country Day School were in court to hear victims describe the shock of having to identify their own faces in pornographic photos to detectives. Juvenile proceedings in Pennsylvania are normally closed, but this was opened by the judge, providing an unusual opportunity for the community to be seen and heard.

The girls described the fallout — anxiety attacks, a loss of trust, problems focusing on schoolwork and a fear that the images may someday surface in unexpected ways.

The two defendants stood stone-faced throughout, flanked by their lawyers and parents, as they were called pedophiles, “sick and twisted” and perverted.

“I will never understand why they did this,” one victim told Judge Leonard Brown, saying it “destroyed my innocence.”

One teen told Brown “how excruciating it is to bring these feelings up again and again.” Another choked back tears as she excoriated one of the defendants for expressing “fake empathy” as girls confided with him about their pain, before it became known that he had been involved. Still another said all of her friends transferred schools, and that she “needed trauma therapy to even walk around my neighborhood.”

Judge said he hadn’t heard boys apologize

The defendants declined several opportunities to comment to the judge, who said he had not heard either boy take responsibility or apologize.

“This has been a regrettable, long, torturous process for everyone involved,” said Heidi Freese, defense attorney for one of the defendants. “There were very interesting, underlying legal issues surrounding the charges in this case and those will be decided on a different day in a different case.”

The other defendant’s lawyers emailed a statement late Wednesday that said he was “extremely remorseful for his part in the AI-generated images and very sorry for any hurt he caused.”

Defense attorneys Adam Szilagyi and Christopher Sarno wrote that the images, which the lawyers said their client did not intend to be public, “contained nudity but did not contain any representations of sexual conduct or activity.”

They said their client did not use “any AI generator himself nor did he disseminate any of the images.” Szilagyi said in a follow-up text that his client was accountable as part of the conspiracy and that both of the boys “gathered and exchanged the unaltered/original images that were put into the generator.”

Brown ordered each to perform 60 hours of community service, have no contact with the victims and pay an unspecified amount of restitution. If they don’t have any additional legal problems, Brown said, the case can be expunged after two years.

As he imposed his sentence, Brown said that if they were adults, they probably would be headed for state prison. He said they should “take this opportunity to really examine” themselves.

Rise of AI has led to deepfakes

The resolution of the Pennsylvania case comes days after three teenagers in Tennessee sued Elon Musk’s xAI, claiming the company’s Grok tools morphed their real photos into explicitly sexual images. The high school students are seeking class-action status to represent what the lawsuit says are thousands of people who were similarly victimized as minors.

The scandal in Pennsylvania led to a student protest, criminal charges against the two teenagers and the departure of leaders at the school, which says it has about 600 students K-12, class sizes averaging just 12 kids, and “an endowment in excess of $25 million.”

Nadeem Bezar, a Philadelphia lawyer who represents at least 10 of the victims, said Tuesday he expects to file a claim “against the school and anybody else we think has culpability in these deepfakes being created and disseminated.”

He said he has not yet seen the photos but expects the legal process to determine “exactly when and where and how the school knew, how the boys created these images, what platforms they used to create these images and how they were disseminated.”

As AI has become accessible and powerful, lawmakers across the country have passed laws aimed at barring deepfakes.

President Donald Trump signed the Take it Down Act last year, making it illegal to publish intimate images including deepfakes without consent, and requiring websites and social media sites to remove such material within 48 hours of being notified by a victim.

Forty-six states now have laws addressing deepfakes, with legislation introduced in the remaining four — Alaska, Missouri, New Mexico and Ohio — according to the consumer advocacy group Public Citizen. ___

Associated Press writers Geoff Mulvihill in Haddonfield, New Jersey, and Holly Ramer in Concord, New Hampshire, contributed.