AI Exposure and Teen Mental Health: Confronting Deepfakes and Chatbot Addiction

For young people just entering the labyrinth of hormones, heartache, discovery, and the search for identity that's adolescence, AI can be more than innovative technology. It can become an ever-present friend that a young person bonds so closely to that it becomes an addiction.

It can also be used to violate their sense of safety and well-being through deepfake sexualization.

Deepfakes are fabricated images and audio tracks created using AI from actual pictures, videos, and voice recordings. Perpetrators can take anyone’s image and/or voice from any source and create a digital likeness. The transgressor can make that image move, speak, even dress or undress, any way they want. And anyone can do it.

“Today's deepfakes use generative AI models, such as GANs or diffusion, trained on a target's images, video, or voice,” according to George Gerchow, IANS  Faculty & Chief Security Officer at Bedrock Data. “Real-time face-swap for live video calls now runs on consumer hardware, and voice cloning needs as little as three seconds of clean audio.

“The barrier has collapsed in the last 18 months. Anyone with a laptop can now do what only well-resourced threat actors could do two years ago.”

Altering images to sexualize children has become an international concern. UNICEF, the United Nations agency working to protect children, reported on the issue earlier this year.

“New evidence confirms the scale of this fast-growing threat: In a UNICEF, ECPAT, and INTERPOL study across 11 countries, at least 1.2 million children disclosed having had their images manipulated into sexually explicit deepfakes in the past year,” according to a statement issued by UNICEF.

Up to two-thirds of children surveyed said they worry that their likenesses could be transformed by AI into fake sexual images or videos, according to the study.

“We must be clear,” the statement reads. “sexualised images of children generated or manipulated using AI tools are child sexual abuse material (CSAM). Deepfake abuse is abuse, and there is nothing fake about the harm it causes.

“When a child's image or identity is used, that child is directly victimized.”

Dr. Jacquelyn Flood, PsyD, is a Child and Adolescent Psychologist at Emora Health. She also dealt with trauma treatment during her service in the Army. 

Being the subject of a non-consensual deepfake, she says, can be “very similar to a trauma-type experience where they feel like they have become vulnerable in a way that’s not even them.”

A few people may respond to the use of their image in deepfakes with outrage, according to Flood.

“Others will retreat into themselves,” she says, “and think, ‘the world is out to get me, and I can’t do anything about it. I’m a victim.’”

Technology is usually ahead of the law, but the two are drawing closer.

The first U.S. law governing non-consensual sexualized deepfakes was signed in May 2025. The Take It Down Act prohibits the creation or distribution of unauthorized explicit deepfakes. Violators face up to two years in prison for cases involving adults. For offenses involving minors (children under 18), the maximum penalty is three years.

Most states have updated or expanded anti-pornography laws to cover deepfakes.

In addition, tech giants and related companies have formed the Coalition for Content Provenance and Authenticity (C2PA) to combat misinformation and deepfakes.

“Watermarking, C2PA content credentials, and platform-level disclosure are where this fight is actually won,” according to Gerchow.

“Bans without provenance are unenforceable because the supply of generation tools is global and open-source. If you can prove what is real, you can isolate what is fake.”

These laws provide ways for victims of deepfakes to counter such attacks. 

• The National Center for Missing and Exploited Children (NCMEC) provides a platform for removing images taken when the subject was a minor.
• The Take It Down Act has established guidelines requiring social media platforms to remove sexualized deepfakes once they're reported.
• The Cyber Civil Rights Initiative (CCRI) provides legal resources and crisis support for victims of pornography and image-based sexual abuse.

In addition to Deepfake exploitation, AI chatbot interactions can lead to addiction. Both outcomes may be tied to reduced human interaction, therapists tell Safewise. That lack of person-to-person contact can reduce empathy in deepfake perpetrators and fuel alienation in chatbot addicts.

At the same time technology is advancing, there is a documented decline in human conversation.

Researchers from the University of Missouri-Kansas City and the University of Arizona have determined that, on average, we spoke 338 fewer words per day each year between 2005 and 2019. That's a decrease of 28 percent. The decline among people under twenty-five was even greater, at 44 percent.

“This loss of words reflects real loss of conversations,” wrote study authors Valeria A. Pfeifer, PhD, and Matthias R. Mehl, PhD, “big ones and small ones, that we stopped having with others.”

Pfeifer told UMKC Today, “Speaking less means spending less time connecting with others. Research has consistently linked loneliness with negative outcomes for both mental and physical health. At the same time, everyday conversation is associated with greater well-being.” 

Therapists Safewise interviewed agree parental communication is the key to helping children who have been abused through deepfakes. However, establishing that communication should start early.

“It’s important to make sure you’re figuring out ways to listen and hear their thoughts and feelings,” says Flood.

To encourage more communication, Flood suggests asking open-ended questions about topics that interest your child.

“I think if you figure out what kinds of things they are more comfortable talking about, it allows you to build your relationship,” said Flood. “Sometimes parents think their child is into XYZ hobby and think, ‘I don’t know anything about that’, but you don’t need to know anything about it.”

The same openness applies to talking to children with an addiction to chatbots, according to Dr. Maya Reynolds, MD, MPH, Psychiatrist at ChoicePoint Health.

“To have a productive conversation with your teen, I strongly advise parents to keep their focus on curiosity only instead of expecting confrontation,” notes Reynolds. “Parents should talk about how their teen feels about talking to a chatbot. This will create space for a healthy conversation without making teens defensive.”

“Research is coming out showing kids have fewer social skills, fewer friendship skills,” says Gayle Alexander, licensed professional counselor and owner/clinical director of Project SAFE. “Language and linguistic skills are down; things like reading a room and understanding a conversation are all down, because they’re doing everything through a screen.”

What begins innocently can grow into trouble.

“A teen may initially engage a chatbot out of curiosity, boredom, loneliness, or looking for stress relief,” contends Dr. Lance Garrison, Dean of the College of Professional Psychology at The Chicago School. “Over time, however, the interaction can begin to feel emotionally significant because the chatbot is always available, highly responsive, remembers personal details, and adapts to the teen’s emotional style.”

The illusion of trustworthiness also factors into chatbot addiction, according to Alexander.

“Most kids do realize that a chatbot is not a real friend,” says Alexander, “but it is somebody who is available 24/7, is never going to spread a rumor about you, is always going to be supportive and tell you what you want to hear. For a generation of kids and young adults whose main source of communication has been texting and social media, this feels normal.”

That feeling of normality can be harmful, according to Garrison.

“Many teens see these systems as private, emotionally safe, and incapable of betrayal. That perception can encourage disclosures that are far more intimate than what they might share with peers, parents, or even therapists,” contends Garrison.

“However, these chats are not confidential, nor are they approximations of therapeutic relationships, which have legal and ethical requirements for maintaining confidentiality. Conversations may be stored, analyzed, or vulnerable to future security breaches or data misuse.”

“Parents can recognize warning signs in ways like observing emotional distress in their teens when they are not able to access a chatbot, withdrawal from family, friends, and activities, keeping secrets about conversations from parents, disrupted sleep routine, and preferring AI guidance over in-person conversation,” according to Reynolds.

Every counselor Safewise interviewed advised against immediately cutting off chatbot interaction. Instead, they recommended an open, non-threatening conversation and a weaning process that replaces screen time with interpersonal activities. In some cases, professional mental health support may be necessary.

“First and foremost, intervention begins with parent engagement. Non-judgmental, genuine, and gentle conversation about the teen’s experience is far more likely to be effective than punishment, shame, or immediate restriction,” notes Garrison.

“The goal is not simply to eliminate technology, but to strengthen human relationships, coping strategies, and support systems that may currently feel less accessible or emotionally rewarding to the adolescent.”