Educators Lack Deepfake Policies and Training

A nationwide EdWeek Research Center survey of 852 educators found that one in three do not know whether their school or district has a policy addressing student safety risks from AI-generated deepfakes. Another 22 percent said their school or district has no such policy, and 2 percent said officials are considering one. The survey was conducted this summer. A separate 2024 survey by the Center for Democracy and Technology found that one in four K-12 students know someone depicted in a nonconsensual, sexually explicit deepfake.

Free-to-download AI tools let students make and distribute fake images or videos of classmates and educators with little effort, the article reports. Many are sexually explicit and target women or girls; others portray school staff making racist or inappropriate remarks. Bart Caylor, founder of school crisis-response marketing firm Caylor Solutions, said creating one can take only a video and a simple instruction to alter what a person says or does.

Most educators surveyed had received no training on identifying or responding to deepfakes, and 13 percent said they did not know what a deepfake is. Nine in 10 said they had received no training on detection or response in the past two years. InnovateEDU chief executive Erin Mote said teachers may be the first to encounter explicit images circulating among students and recommended a unified incident-response plan to help staff detect and report cases safely, limit distribution and support victims with trauma-informed care. She warned teachers not to send intimate images of minors to personal phones, which could put child sexual abuse material on the device.

More than 20 percent of respondents said existing school policies were broad enough to cover sexually explicit deepfakes. However, experts said such cases differ from cyberbullying: distributing a nonconsensual sexually explicit image of a minor is a federal crime, and some state measures and a federal law include jail time as a possible consequence. Kristin Woelfel of the Center for Democracy and Technology said schools can amend existing technology-use, harassment and cyberbullying policies with specific language rather than necessarily creating new policies. She said responding to harm matters more than determining whether an image is real, because the effect on victims is the same.

The survey found that 56 percent of schools would most likely inform parents when a student creates or distributes a harmful deepfake; 48 percent would suspend the student, a quarter would expel them, and 33 percent would use restorative practices. Mote said students need plain-language explanations of school rules and consequences, while Woelfel said schools should teach consent and tell students that creating or sharing such material is wrong. Separately, lawmakers in 28 states have adopted laws regulating AI-generated deepfakes in political messaging.

A nationwide EdWeek Research Center survey of 852 educators found that one in three do not know whether their school or district has a policy addressing student safety risks from AI-generated deepfakes. Another 22 percent said their school or district has no such policy, and 2 percent said officials are considering one. The survey was conducted this summer. A separate 2024 survey by the Center for Democracy and Technology found that one in four K-12 students know someone depicted in a nonconsensual, sexually explicit deepfake.

Free-to-download AI tools let students make and distribute fake images or videos of classmates and educators with little effort, the article reports. Many are sexually explicit and target women or girls; others portray school staff making racist or inappropriate remarks. Bart Caylor, founder of school crisis-response marketing firm Caylor Solutions, said creating one can take only a video and a simple instruction to alter what a person says or does.

Most educators surveyed had received no training on identifying or responding to deepfakes, and 13 percent said they did not know what a deepfake is. Nine in 10 said they had received no training on detection or response in the past two years. InnovateEDU chief executive Erin Mote said teachers may be the first to encounter explicit images circulating among students and recommended a unified incident-response plan to help staff detect and report cases safely, limit distribution and support victims with trauma-informed care. She warned teachers not to send intimate images of minors to personal phones, which could put child sexual abuse material on the device.

More than 20 percent of respondents said existing school policies were broad enough to cover sexually explicit deepfakes. However, experts said such cases differ from cyberbullying: distributing a nonconsensual sexually explicit image of a minor is a federal crime, and some state measures and a federal law include jail time as a possible consequence. Kristin Woelfel of the Center for Democracy and Technology said schools can amend existing technology-use, harassment and cyberbullying policies with specific language rather than necessarily creating new policies. She said responding to harm matters more than determining whether an image is real, because the effect on victims is the same.

The survey found that 56 percent of schools would most likely inform parents when a student creates or distributes a harmful deepfake; 48 percent would suspend the student, a quarter would expel them, and 33 percent would use restorative practices. Mote said students need plain-language explanations of school rules and consequences, while Woelfel said schools should teach consent and tell students that creating or sharing such material is wrong. Separately, lawmakers in 28 states have adopted laws regulating AI-generated deepfakes in political messaging.