One in 3 educators don’t know if their school or district has a policy to address the potential threat to student safety from AI-generated fake images and videos.
That’s according to a nationwide survey of 852 educators by the EdWeek Research Center conducted this summer. A smaller share—22%—said their school or district doesn’t have such a policy, and 2% of respondents said their district or school is “considering” creating one.
The problem of what’s known as “deepfakes”—which could falsely feature an educator or student saying something offensive or appearing in a sexual manner—is growing in schools at a much faster pace than the enactment of policies to address their harm.
Nationally, 1 in 4 K-12 students know someone who has been depicted in a nonconsensual and sexually explicit deepfake, according to a 2024 student survey by the Center for Democracy and Technology, a nonprofit that advocates for people’s rights online.
The internet is now rife with simple and free-to-download AI tools that students can use to create and distribute fake images or videos of their peers or educators. Many are sexually explicit, targeting women or girls. Others threaten to damage reputations—for example, students have made fake videos of principals or other employees making racist or inappropriate remarks.
“Educators may not realize how little skill is needed to do this,” said Bart Caylor, the founder of Caylor Solutions, a marketing firm that works with schools to create crisis response plans. “It’s no more than grabbing a video from Facebook, throwing it into one of these tools and saying, ‘Change what they’re saying and doing.’ It doesn’t take any skill, it doesn’t take any effort, it doesn’t take any money.”
School and district leaders should put together a plan of action before an incident occurs, rather than scrambling after the fact. However, that could be challenging if educators aren’t aware of deepfakes or the damage they can cause.
In the EdWeek Research Center survey, 13% of educators said they didn’t know what a deepfake is, and most respondents hadn’t received any training on spotting or responding to deepfakes.
Such training is especially important for teachers, who are likely the first line of defense when an explicit deepfake starts making the rounds among students, said Erin Mote, the founder and chief executive officer of InnovateEDU, a national education nonprofit that helps school districts create safety guidelines against the improper use of AI by students.
“I think leaders need to know: Are their educators clear about what a deepfake is and what to do when they encounter it?”
Can existing discipline policies work?
More than 20% of respondents in the survey said their school or district’s existing policies are broad enough to cover sexually explicit deepfakes.
But a deepfake case isn’t the same as cyberbullying, experts concur. While students may use AI to generate images with the intent to bully, publishing and distributing a nonconsensual, sexually explicit image of a minor is now a federal crime, which allows law enforcement to penalize perpetrators who are minors themselves.
These AI-generated images or videos can also cause irreparable damage to young people who have no control over how their likeness is manipulated and shared. When deepfakes surface in school, student victims often have to sit in class with those who made or circulated the deepfake, and appeal directly to social media platforms to get these images taken down.
Deepfakes need to be a bigger part of the conversation on student safety, but that doesn’t mean districts necessarily need a new policy, said Kristin Woelfel, policy counsel at the Center for Democracy and Technology.
Woelfel, who has worked on model policies to address deepfake incidents in schools, said it’s important to amend or add specific language within existing policies about fair technology use, harassment, and cyberbullying to address the harm caused by these media.
“Schools can be creative about where they put it,” said Woelfel. “I think that it’s even part of prevention. Not everybody would know deepfakes fall under these policies because we haven’t been dealing with this on a wide scale. But at this [current] scale, schools are likely to see more deepfakes.”
How leaders respond to deepfakes is critical
Schools need to draft an “incident response plan,” which can borrow from existing policies, but is a one-stop, unified approach for everyone to follow when a deepfake surfaces, said Mote.
The plan should help educators detect and report deepfake incidents safely, curtail their spread, and support victims with trauma-informed care, Mote added. Teachers should be trained to spot the signs of a fake image and follow established protocols—like mandated reporting—when a student brings an inappropriate image to their attention.
“The last thing you want is for a teacher to say, ‘Hey, send me that deepfake intimate image to my personal cellphone so I can take it to the principal,’” Mote said. “Because now the educator has access to child sexual abuse material on their personal phone.”
Schools can offer professional development for teachers on AI-generated media to build their knowledge on detection and ensure they’re following the right protocols, Mote added.
For Woelfel at the Center for Democracy and Technology, detection isn’t as important as responding in the right way. For sexually explicit images that are being shared without consent, research shows the harm to a victim “is ultimately the same, whether it’s real or fake,” she said. Figuring out whether an image is real “is actually less important than figuring out what you’re gonna do about it.”
Overall, though, schools and districts are not training educators to detect or respond to deepfakes. Nine in 10 educators told the EdWeek Research Center that they haven’t been trained on either topic in the last two years.
Prevention is about making sure students know the consequences
Students need to be fully aware of their school’s deepfake policies or response plans that clearly state consequences. Mote said school districts “owe students some plain language” about what could happen if they create or distribute nonconsensual images or videos; some state measures and a federal law include jail time as one of the consequences.
At present, schools are most likely (56%) to inform parents if a student makes or distributes a harmful deepfake, according to the survey. Another quarter said they’d expel student perpetrators, 48% would use suspension as a punishment, and 33% said they’d use restorative practices to deal with the fallout.
Ideally, students would learn to be more empathetic to their peers and develop a deep understanding of consent, Woelfel said.
“At the very least, schools should tell people that this is wrong,” said Woelfel. “If you don’t do that, then you’re not even one step forward toward that process.”
