Nearly half of teens have already seen artificial intelligence (AI)-generated explicit sexual material, and almost a quarter have seen it depicting themselves or someone they know, according to new research from Common Sense Media’s Youth AI Safety Institute.
Across the country, local educational agencies, teachers and students have reported increases in cyberbullying and other issues associated with sexually explicit AI-generated materials.
According to Common Sense Media, “AI has lowered the barrier for access and exposure to this content by streamlining the need for technical skills to create sexualized content of a real person, without needing to access the person or receiving consent. And while the majority of young people believe this should be illegal, there’s a growing gap between that belief and the reality of how this content is easily created, manipulated, and shared.
“Technologies that can generate sexual content are too easily accessible to teens, and the social media platforms that allow the distribution of this content are not doing a good enough job of detecting, blocking, and taking it down,” the report continues. “This research reveals not only the scale of this exposure and the depth of the impact, but also the degree to which teens are navigating this largely alone. They are encountering content that makes them worry about being a target, they are watching peers get hurt by the manipulation of their image, and they are developing their own frameworks around consent, bodily ideals, and online harm — often without the guidance of a parent, caregiver, or teacher. Our research suggests this can lead to feelings of helplessness and a lack of agency in young people because they feel they can do nothing about their experiences.”
Common Sense Media founder and CEO James P. Steyer said in a statement that parents and guardians are largely unaware of how prevalent the issue is.
Key findings
More than 1,300 teens aged 13-17 across the U.S. were surveyed for this report. Their responses show that:
- Sixty-seven percent of teens fear being a victim of a deepfake without consent, and many are scaling back online activity to protect themselves. Their actions include making social media accounts private (30 percent), posting images of themselves less (23 percent) or deleting social media accounts (9 percent). Girls are more worried than boys of being a target (71 percent and 62 percent, respectively).
- Eighteen percent report having made AI pornography or knowing someone who has, and 25 percent have shared it or know someone who has. Among those who have created AI pornography or know someone who has, 63 percent have created pornography depicting themselves or people they know. Boys are almost three times as likely as girls to have made it (11 percent compared to 4 percent).
- While 82 percent of teens who report having seen AI-generated pornography say creating it without consent should be illegal, 23 percent believe AI nudes are less harmful than real ones “because no one gets hurt.”
Recommendations
The report notes that addressing the challenges presented by this type of AI content requires action on multiple fronts. “Platforms need better tools for detection and reporting of this type of content. Lawmakers need to pass legislation with robust enforcement of penalties against the AI tools and platforms that allow such content to be generated and spread. Parents and caregivers need resources that help them start conversations. And schools need policies and curricula that reflect the digital environments of today’s teens,” the report states.
Recommendations for LEAs include:
- Teach students how AI-generated sexual content spreads and who it can affect. Many teens encounter it without any framework for understanding the real-world impact of this content.
- Teach students why it’s harmful, including that AI-generated content can violate consent and cause real harm, even when no one appears directly involved.
- Talk about this topic before a crisis hits. Students are more likely to come forward when they’ve already experienced an educator engaging with it openly and without judgment. Make it clear that students who encounter or are targeted by this content won’t be judged for it. Fear of blame is one of the top barriers that teens identify to coming forward.
- When making procurement decisions, ensure that any AI tools used by schools are incapable of generating sexual or intimate images, and tighten school-issued device filters so that unauthorized AI apps and websites are not available to students.
- Some students will need support as targets or bystanders while others need accountability and education as creators or sharers. Effective school response addresses both without collapsing them into one conversation.
Learn more about this issue in the spring edition of California Schools, which includes a deep dive into the challenges students and LEAs are facing related to sexploitation, deepfakes and more.
