When AI Becomes a Weapon of Digital Peer Aggression: The Growing Use of Non-Sexual Deepfakes to Humiliate, Embarrass, and Bully Youth and Teens
CAVEAT: Here at The White Hatter, we have written extensively about the growing concern surrounding artificial intelligence (AI) being used to create non-consensual intimate images, commonly referred to as sexual deepfakes. However, we believe there is another dimension of AI generated content that deserves greater attention from parents, caregivers, and educators, the use of AI generated images, videos, and audio recordings that have nothing to do with nudity or sexual exploitation, but are instead created to embarrass, ridicule, impersonate, intimidate, or socially isolate another person. These non-sexual deepfakes can become powerful tools of digital peer aggression, particularly within school communities where friendships, social acceptance, reputation, and belonging often play an important role in adolescent development. Although the technology may be relatively new, the underlying behaviours are not. Bullying, humiliation, impersonation, and social exclusion have existed for generations. What has changed is the ability to manufacture convincing visual or audio content depicting someone doing or saying something that never actually happened, and then distribute that fabricated content to potentially hundreds or thousands of people within minutes.
We also want to acknowledge that not every AI generated face swap, altered photograph, synthetic voice, or manipulated video is inherently harmful. There are legitimate creative, educational, artistic, and entertaining uses for this technology, particularly when the people involved have provided informed consent and the resulting content is presented transparently. Many youth and teens are experimenting with these tools out of curiosity, creativity, and a desire to participate in emerging digital trends. We do not believe that experimentation should automatically be characterized as dangerous or inappropriate. The concern arises when a person’s identity, likeness, or voice is used without their knowledge or permission in circumstances intended to deceive, humiliate, or cause harm. Understanding this distinction is important because the solution should not be to frighten young people away from AI technology, but rather to help them develop the judgment, empathy, character, and digital literacy skills necessary to use it responsibly.
There was a time when creating a convincing deepfake required specialized software, considerable computing power, technical knowledge, and a relatively sophisticated understanding of artificial intelligence, that is no longer the case. Today, consumer facing AI applications have made it increasingly easy for everyday users to manipulate existing photographs and videos, replace faces, generate synthetic voices, and create fictional situations involving real people. Applications such as CapCut, Kling AI, and other generative AI image and video platforms have helped make sophisticated digital manipulation accessible to a much broader audience, including youth and teens who may have little or no formal training in video production or artificial intelligence. Depending on the application and its available features, a user may only need to upload a photograph, select an existing video, or enter a written instruction describing what they want the AI to create.
One type of content circulating on social media involves taking an existing video and using AI to replace the person appearing in that video with someone else. Depending on the technology being used, the altered version can preserve much of the original video’s movement, timing, background, and physical actions while changing the identity of the person depicted. A creator might insert themselves into a popular dance routine, replace an actor in a movie scene, or place a celebrity into a humorous viral video. In many cases, the result is intended to be entertaining, and viewers understand that the content has been artificially generated. However, the same process can be used to take a photograph of an unsuspecting classmate and insert their likeness into a fabricated situation that is embarrassing, degrading, offensive, or socially damaging.
Consider how easily a student’s photograph can now be obtained. Their image might appear on a school website, in a sports team photograph, on a public social media account, in a yearbook, or within a group chat. Depending on the AI application, that photograph may provide enough visual information to create a synthetic image or video depicting the student participating in an event that never occurred. The student may have no knowledge that their photograph has been copied, uploaded into an AI platform, or used to create manipulated content until the resulting image or video begins circulating among classmates. In some cases, the fabricated content may look sufficiently realistic to deceive viewers. In other cases, the manipulation may be obvious, but the content can still become a vehicle for humiliation because the objective was never necessarily to convince people that the event actually happened.
This is one of the most important distinctions we believe parents, caregivers, and educators need to understand. A harmful deepfake does not have to be believable to be effective. Sometimes, the entire purpose is to make someone look ridiculous, invite others to laugh at them, or reinforce their position as an outsider within a peer group. The technology simply provides another means of achieving that objective.
Digital Peer Aggression: An Old Behaviour With a New Technological Capability
For years, researchers and educators have examined how digital technology can be used to facilitate cyberbullying, harassment, humiliation, impersonation, exclusion, rumour spreading, and other forms of electronic aggression (1). Artificial intelligence has not invented these behaviours, and we shouldn’t assume that every young person who experiments with AI generated content intends to harm someone. What AI has introduced is the ability to manufacture visual and audio material that can make an existing act of peer aggression more convincing, more personalized, easier to reproduce, and potentially more difficult for the targeted individual to challenge.
This is not simply a theoretical concern about what artificial intelligence might allow young people to do in the future. Here at The White Hatter, we have already documented how AI is changing the nature of digital peer aggression among youth and teens. In our article, “How AI Is Changing Digital Peer Aggression (Cyberbullying) for Today’s Youth and Teens” (2), we discussed how students are increasingly able to manufacture convincing images, videos, and audio recordings depicting events that never occurred. We also shared how a teen had contacted us after experiencing AI driven peer aggression, reinforcing that this technology is already being used to target young people. What makes this development particularly concerning is that traditional cyberbullying often involved sharing, manipulating, or weaponizing something that actually existed, such as a photograph, message, or embarrassing incident. Generative AI changes that dynamic because the person engaging in the aggression no longer requires an authentic event to exploit. They can manufacture the event itself and then use that fabrication to influence how others perceive, ridicule, or treat the person being targeted.
Imagine a hypothetical situation involving a group of students who dislike one of their classmates. In the past, they might have spread an embarrassing rumour, created an insulting meme, altered a photograph using conventional editing software, or posted hurtful comments in a group chat. Today, one of those students could use an AI application to generate a video depicting that classmate making an offensive statement, behaving inappropriately, crying uncontrollably, or participating in an embarrassing activity. The video could then be shared within a private group chat where other students begin laughing, commenting, adding captions, and forwarding it to additional groups. Within a relatively short period, the fabricated content could circulate throughout a school community, reaching students who may have no understanding of how it was created or whether the depicted event actually occurred.
For the student being targeted, the distinction between a genuine recording and an AI generated fabrication may provide little emotional comfort. Some classmates may believe the content is authentic and begin treating the student differently because of what they think they witnessed. Others may recognize that the video is fake but continue sharing it because they find the humiliation entertaining. The targeted student may then find themselves having to explain that something never happened, defend their reputation against a fabricated event, or endure ongoing ridicule over content they had no involvement in creating. This can be especially challenging during adolescence, when social belonging, peer acceptance, friendships, and personal identity are often closely connected.
We believe this represents an important evolution in digital peer aggression. Traditionally, a bully might attempt to damage another person’s reputation by spreading a false story about them. With AI, that same individual may now be able to create fabricated visual or audio material that appears to support the false story. In other words, the technology can potentially transform a rumour into something that looks like evidence, even though the underlying event never occurred.
Not Every Act of Digital Meanness Is Bullying, But That Does Not Mean It’s Harmless
It is important to recognize that not every act of digital meanness, conflict, or inappropriate humour automatically meets the definition of bullying. As we discussed in our article, “Cyberbullying & Online Aggression” (3), understanding the distinction between interpersonal conflict, isolated inappropriate behaviour, and sustained peer aggression requires consideration of several factors, including intent, repetition, power imbalances, and the resulting harm. We believe these distinctions matter because the response to a single impulsive act of poor judgment may be very different from the response required when a student is being deliberately and repeatedly targeted by someone seeking to gain power, influence, or social status within a peer group.
Artificial intelligence does not eliminate the importance of these distinctions, but it can complicate them. A single AI generated image or video may be repeatedly distributed, modified, and weaponized by multiple students, even when the original creator only produced the content once. What initially appears to be an isolated incident can therefore become part of a broader pattern of humiliation or harassment, particularly when the targeted student has limited ability to stop its circulation or challenge the social influence of those distributing it. In some circumstances, the repeated sharing of the same fabricated video may become more harmful than the original act of creating it.
This is also why we believe parents and educators need to be cautious about automatically labelling every inappropriate AI generated image as bullying. A student who creates an altered photograph of a friend without considering how it might be received may have exercised poor judgment, but that does not necessarily mean they intended to intimidate or socially isolate that person. However, if the student continues distributing the content after being asked to stop, deliberately encourages others to ridicule the person depicted, or uses the material as part of an ongoing pattern of aggression, the circumstances become very different.
The objective should be to understand the behaviour, identify the harm, and respond proportionately rather than applying the same label or consequence to every incident.
Non-Sexual Deepfakes: The Many Ways AI Can Be Used to Target Another Student
When parents hear the word “deepfake,” their immediate concern often relates to the creation of non-consensual intimate images. This concern is understandable given the serious emotional, social, and legal implications associated with sexual deepfakes. However, focusing exclusively on nudity and sexually explicit content risks overlooking other forms of AI facilitated aggression that can also cause meaningful harm.
In our article, “AI: The Next Challenge in Online Safety, Security, and Privacy When It Comes to Youth and Teens” (4), we explored how artificial intelligence introduces new opportunities for the manipulation of digital identities, including the creation of fabricated images, videos, and audio recordings. Although public discussions about harmful deepfakes have understandably concentrated on sexually explicit content, the same underlying technology can be used to manufacture false allegations, embarrass classmates, damage friendships, or influence how someone is perceived within their school community. We believe that parents, caregivers, and educators who focus exclusively on AI generated nudity may overlook other forms of synthetic media abuse that can also have significant social and emotional consequences.
Humiliation Based Deepfakes:
One example involves what we would describe as humiliation based deepfakes, where a student’s face is placed onto another person’s body or inserted into a fabricated scene designed to make them appear foolish, awkward, unattractive, or socially undesirable. A student might be depicted participating in an embarrassing dance, falling over repeatedly, making exaggerated facial expressions, or behaving in a manner that invites ridicule. Although some of these examples may appear relatively harmless when viewed in isolation, the context in which they are created and distributed matters. If the content is being used repeatedly to target a particular student, reinforce an existing pattern of bullying, or encourage others to participate in public humiliation, the behaviour can become much more serious than a single inappropriate joke.
Reputational Deepfakes:
Another concern involves reputational deepfakes, where AI is used to falsely depict a student making discriminatory statements, insulting teachers, threatening classmates, admitting to misconduct, or engaging in other behaviour that could damage their standing within a school community. Unlike an obviously humorous face swap, this type of fabricated content may be designed specifically to deceive viewers into believing that the depicted event actually occurred. A student could find themselves facing anger, social rejection, disciplinary scrutiny, or damaged friendships because of something they never said or did. This creates an additional challenge for educators, who may encounter what appears to be compelling video or audio evidence of student misconduct without immediately recognizing that the material could have been artificially generated.
Relationship Based Deepfakes:
Relationship based deepfakes present another potential concern, particularly among adolescents navigating friendships, dating, romantic interests, and social hierarchies. AI generated content could falsely depict two students kissing, dating, arguing, or participating in other interactions that never occurred. Although such content may not be sexually explicit, it could still be used to embarrass someone, create jealousy, damage an existing relationship, or provoke conflict between peer groups. For a young person who is already experiencing social difficulties, the circulation of a fabricated relationship video could become particularly distressing, especially if classmates begin treating the fictional event as genuine.
Voiced Based Deepfakes:
We are also concerned about the use of AI generated voice recordings and impersonation. A student’s voice could potentially be imitated to make it sound as though they are insulting a friend, threatening another student, revealing private information, or making statements that conflict with their actual beliefs. In these circumstances, the manipulation may be less visually obvious because the content involves audio rather than video. Given how frequently young people communicate through voice messages, gaming platforms, and private group chats, synthetic audio introduces another opportunity for digital peer aggression that parents and educators may not immediately recognize.
What connects these examples is not the presence or absence of sexual content, but the deliberate manipulation of another person’s identity to influence how others perceive or treat them. The harm comes from the misuse of someone’s likeness, voice, or reputation, not simply from the technical process used to create the content.
When “It Was Just a Joke” Becomes an Excuse for Digital Humiliation
One of the explanations we frequently hear when discussing inappropriate online behaviour involving youth and teens is that the content was created as a joke and that the person being targeted should not take it so seriously. Sometimes, this explanation may reflect a genuine misunderstanding of how the content would be received. Young people often experiment with humour, satire, parody, and social boundaries as part of their development, and not every embarrassing photograph or manipulated video should automatically be classified as bullying. Context, relationships, intent, consent, repetition, and impact all matter when evaluating whether a particular incident represents harmless humour, poor judgment, or deliberate digital aggression.
However, there is an important difference between laughing with someone and deliberately creating content that encourages others to laugh at them. When a student knowingly uses another person’s likeness to manufacture an embarrassing or degrading video without their permission, particularly when the objective is to generate ridicule among classmates, the behaviour deserves closer examination. If the person being depicted expresses discomfort, asks for the content to be removed, or becomes visibly distressed, continuing to distribute the material under the justification that it is simply a joke becomes increasingly difficult to defend.
Parents and educators should also recognize that young people do not always openly communicate when something has hurt them. A student may laugh along with their peers because they do not want to appear overly sensitive, attract additional attention, or risk becoming an even greater target. Others may remain silent because they fear that objecting will encourage the creator to produce additional content. This is particularly relevant within adolescent peer groups, where social pressure and the desire to belong can influence how young people respond to embarrassing situations.
We believe an important question for parents, caregivers, and educators to ask is not simply whether the creator intended the content to be funny, but whether the person being depicted had a meaningful opportunity to consent and whether the resulting content respects their dignity. A student should not have to surrender control over how their identity is represented online simply because someone else believes their humiliation would make entertaining content.
Why Group Chats Can Turn One Deepfake Into a School Wide Incident
One of the realities of today’s youth digital culture is that much of their social interaction now takes place within private or semi private digital environments. Group chats, direct messages, gaming communities, and smaller online social networks have become important spaces where young people communicate, maintain friendships, organize social activities, and develop a sense of belonging. Although public facing social media platforms continue to play a role in adolescent life, private digital communities are often where much of the everyday social interaction between peers now occurs.
Unfortunately, these same environments can also become distribution networks for harmful AI generated content. A student who creates an embarrassing deepfake may initially share it with only a few close friends, believing that the content will remain within that small group. However, recipients may forward the video to other conversations, save copies to their devices, screen record it, or upload it to a public platform. Others may add insulting captions, remix the original video, or use additional AI applications to create new variations. What began as a single manipulated image or video can therefore develop into a broader incident involving multiple students, several digital platforms, and an increasingly difficult effort to contain the material.
We have previously explored how adolescent social hierarchies, peer influence, and group dynamics can contribute to AI facilitated aggression. In our article, “The New Face of Mean-Girl Aggression: AI-Generated Nudes and Digital Harm” (5), we discussed a case involving a teen girl who used AI generated intimate images to target other girls at her school. Although that particular incident involved sexualized imagery, the underlying social dynamics are equally relevant to non-sexual deepfakes. The desire for social status, peer approval, retaliation, power, or control can influence why some young people choose to manipulate another person’s identity and distribute the resulting content. Group chats and private messaging environments can further reinforce these behaviours when recipients respond with laughter, encouragement, or additional sharing. The lesson is that the technology may change, but the human motivations behind peer aggression often remain familiar.
The social dynamics surrounding this behaviour are particularly important. Some students may participate because they genuinely find the content amusing and have not considered its impact. Others may share it because they want to fit in, gain social approval, or avoid becoming the next target themselves. There may also be students who recognize that the behaviour is harmful but remain silent because they do not want to challenge influential members of their peer group. In this way, digital peer aggression can become a collective behaviour in which the original creator is only one part of a much larger social process.
This is why we believe digital citizenship education needs to address not only the person who creates harmful content, but also the responsibility of those who receive, react to, and distribute it. A student does not need to manufacture a deepfake to contribute to the harm it causes. Forwarding a humiliating video, adding mocking commentary, or encouraging others to view it can amplify the original behaviour and increase the number of people exposed to the content.
At the same time, students who choose not to participate can play an important role in interrupting this process. Refusing to forward harmful content, supporting the targeted individual, challenging inappropriate behaviour when it is safe to do so, and reporting serious incidents to a trusted adult are all examples of responsible digital citizenship. These actions may appear relatively small, but within a peer environment where social approval can influence behaviour, they can make a meaningful difference.
The Psychological Impact: Why Knowing Something Is Fake Does Not Necessarily Reduce the Harm
One assumption that deserves closer examination is the belief that AI generated content is somehow less harmful because it does not depict an event that actually occurred. From an adult perspective, it may seem reasonable to suggest that a student should simply explain that the video is fake and move on. However, this response can overlook the social and emotional realities of adolescence, particularly when the fabricated content is circulating among people whose opinions matter to the targeted individual.
A student whose likeness has been used in a humiliating deepfake may experience anxiety about who has seen the content, embarrassment when entering a classroom, concern that others believe the fabricated event actually happened, or fear that additional manipulated content will be created. They may become reluctant to participate in group chats, withdraw from social activities, or feel uncomfortable attending school because they do not know how their peers will react. Even when classmates openly acknowledge that the content is artificial, the ridicule associated with it may continue because the purpose of the video was to make the targeted student an object of amusement.
Adolescence is a developmental period during which identity formation, peer relationships, belonging, and social acceptance can carry considerable emotional importance. Being publicly humiliated through manipulated content can therefore affect more than a student’s immediate mood. Depending on the circumstances, it may influence their confidence, sense of social security, willingness to participate in school activities, and trust in their peers. This does not mean that every embarrassing AI generated image will produce lasting psychological harm, and we should be careful not to make that assumption. However, it does mean that adults should avoid dismissing an incident simply because the material is artificial or because the creator claims that no harm was intended.
There is also an additional dimension involving uncertainty and loss of control. Once a student discovers that someone has used their photograph or voice to create a fabricated video, they may begin wondering what else could be generated using their identity. Could another video appear tomorrow? Could the same photograph be used to manufacture something more offensive? Could a future teacher, employer, or friend encounter the content without understanding that it was fabricated? Although these outcomes are not inevitable, the uncertainty itself may contribute to distress, particularly when the student does not know who created the material or how widely it has circulated.
We believe the appropriate response is to take the young person’s experience seriously while avoiding unnecessary fear or catastrophizing. Some incidents may be resolved quickly through communication, removal of the content, and appropriate school intervention. Others may require more extensive support, particularly when the deepfake is part of an ongoing pattern of bullying or harassment. The response should be guided by the circumstances and the needs of the young person affected, rather than by an assumption that all AI generated content is either harmless entertainment or inherently traumatic.
The Digital Echo: Why Removing the Original Deepfake May Not End the Problem
Another important consideration involves what we at The White Hatter describe as the Digital Echo, the reality that content shared online can sometimes persist, be replicated, transformed, or redistributed beyond the original creator’s intentions. When an AI generated image or video is shared within a group chat, recipients may save it, capture screenshots, record their screens, or forward the material to other conversations. Even if the person who created the original deepfake later deletes it, copies may remain on other devices or accounts. In some circumstances, the content may also be uploaded into additional AI applications and modified further, creating new versions that extend the life of the original incident.
This concern is closely connected to something we explored in our article, “From Digital Footprint to Digital Echo: Why AI Is Changing How We Talk About Our Online Identity” (6). For years, digital literacy education has emphasized the importance of managing our digital footprint, meaning the information we personally create, share, or leave behind through our online activities. However, generative AI has introduced another dimension that we believe requires greater attention. A young person can now become associated with images, videos, statements, or behaviours that they never created, participated in, or consented to. Their digital identity can be influenced not only by what they choose to share, but also by what others create about them using artificial intelligence.
This is why we believe the concept of a Digital Echo is increasingly relevant. Information can be copied, replicated, transformed, manipulated, and redistributed in ways that extend beyond the original content. For a student targeted by a non-sexual deepfake, the concern is not simply that someone has created an embarrassing video, but that a fabricated representation of their identity may continue influencing how others perceive them even after the original content has been removed.
This ability to replicate and transform content introduces an additional challenge for families and schools attempting to address digital peer aggression. Removing the original video from one platform does not necessarily eliminate copies that have already been distributed elsewhere. A student may also discover that the manipulated content has resurfaced weeks or months after the original incident, potentially renewing embarrassment or conflict that they believed had been resolved.
However, we also want to be careful not to exaggerate the likelihood of permanent or widespread circulation. Not every embarrassing deepfake becomes viral, and many incidents remain confined to relatively small peer groups. Early intervention, appropriate reporting, and cooperation between students, families, schools, and platforms can often help limit further distribution. The important lesson is not that every digital mistake will follow a young person forever, but that sharing content can create consequences that are difficult to fully control once other people possess copies.
This is why prevention and informed consent matter. Before creating or sharing an AI generated image involving another person, young people should understand that they may not be able to control where that content travels, how others interpret it, or whether someone else decides to modify it further.
The Canadian Legal Context: Not Every Harmful Deepfake Is Automatically Criminal
For Canadian parents, caregivers, and educators, it is important to understand that the legal implications of creating or distributing a non-sexual deepfake depend heavily on the circumstances. There is no single provision in Canada’s Criminal Code that automatically makes every non-consensual AI generated image or video a criminal offence. The mere act of creating a face swap using another person’s publicly accessible photograph does not necessarily meet the legal threshold for criminal prosecution. However, the behaviour surrounding the creation, intended use, and distribution of that content may engage existing criminal or civil laws.
For example, if AI generated material is used as part of repeated threatening or harassing behaviour, certain Criminal Code provisions relating to criminal harassment may become relevant, provided the necessary legal elements are established. Criminal harassment under section 264 requires more than simply causing embarrassment or emotional upset. Its statutory elements include prohibited conduct and a reasonable fear for safety in the circumstances. If a fabricated video or audio recording is used to communicate threats, extort someone, or facilitate fraud or identity related offences, other criminal provisions may potentially apply. However, criminal identity fraud under section 403 also has specific requirements and should not be interpreted as making every act of online impersonation illegal. Similarly, deliberately circulating false material that seriously damages another person’s reputation could, depending on the facts, raise civil concerns involving defamation or other recognized legal remedies.
It is also important to distinguish non-sexual deepfakes from offences involving intimate images. Canada’s criminal provisions concerning the non-consensual distribution of intimate images contain specific legal requirements, and an ordinary non-sexual face swap does not automatically fall within those provisions simply because it was created without permission. The legal analysis can become more complicated when synthetic content depicts nudity or sexual activity, which is one reason we believe non-sexual deepfake aggression deserves to be examined as a distinct issue rather than automatically grouped together with intimate image abuse.
For schools, however, the absence of a criminal offence does not necessarily mean that no intervention is possible. School codes of conduct, bullying prevention policies, and applicable education legislation may provide mechanisms for addressing harmful student behaviour, including certain incidents that occur outside school hours or on personal devices when there is a sufficient connection to the school environment. A student who deliberately creates and circulates a humiliating deepfake of a classmate may therefore face appropriate school based consequences even when police determine that the circumstances do not justify criminal charges.
We believe schools should approach these situations with fairness, proportionality, and an understanding of the technology involved. This includes carefully assessing whether the content is authentic, considering the intent and impact of the behaviour, distinguishing between those who created and distributed the material, and ensuring that the rights and circumstances of all students involved are respected. The objective should be to address harmful behaviour, support the targeted student, and reinforce expectations surrounding responsible technology use rather than simply treating every AI related incident as a disciplinary problem.
What Parents and Caregivers Need to Be Teaching Their Children
Here at The White Hatter, we have consistently advocated that raising children in today’s online world requires more than simply installing parental controls, restricting access to applications, or delaying access to certain technologies. Although these approaches may have a place depending on the age, maturity, and circumstances of the child, they do not replace the importance of developing character, empathy, judgment, critical thinking, and responsible decision making. The increasing accessibility of AI generated deepfakes reinforces why these human skills remain central to digital literacy.
One of the most important conversations parents and caregivers can have with their children involves the concept of consent as it relates to another person’s digital identity. Many young people understand that they should not share someone’s private photograph or personal information without permission, however they may not immediately recognize that using a publicly accessible image to create a fabricated video can raise similar ethical concerns. A photograph appearing on a school website, sports team page, or public social media account does not automatically mean that the person depicted has agreed to have their face inserted into an embarrassing or misleading situation. The fact that an AI application makes this technically possible does not mean that doing so is respectful or appropriate.
Parents and caregivers should also help their children understand the difference between intention and impact. A young person may genuinely believe that a face swap or manipulated video is funny, particularly when their friends are laughing and encouraging them to share it. However, responsible digital citizenship requires considering how the person being depicted might feel, whether they would be comfortable having the content distributed, and whether the resulting image or video could affect their relationships or reputation. These conversations should encourage empathy rather than simply focusing on the possibility of punishment.
Another important lesson involves recognizing that sharing harmful content can contribute to the original behaviour. Young people should understand that they do not need to create a humiliating deepfake to become part of the problem. Forwarding it to friends, posting it in another group chat, adding insulting captions, or encouraging others to view it can increase the harm experienced by the targeted person. Conversely, choosing not to share the content, supporting the person being targeted, and seeking assistance from a trusted adult can help interrupt the cycle of digital humiliation.
We also believe parents should encourage their children to question the authenticity of unusual or controversial content involving people they know. If a video suddenly appears showing a classmate making an offensive statement, behaving inappropriately, or participating in an embarrassing event, young people should be encouraged to consider whether the material could have been manipulated before reacting or sharing it. This does not mean teaching children to distrust everything they encounter online, but rather helping them develop the habit of pausing, questioning, and verifying when something appears unusual or potentially harmful.
Most importantly, these conversations should take place before an incident occurs. Waiting until a child becomes the creator, distributor, or target of a harmful deepfake means that an important opportunity for prevention may already have been missed.
What Educators and Schools Should Be Considering
Schools are increasingly being asked to navigate the challenges created by generative AI, and we believe the growing accessibility of synthetic images, videos, and audio recordings presents both a responsibility and an educational opportunity. For years, digital citizenship education has focused on cyberbullying, privacy, online reputation, responsible communication, and the importance of thinking before posting. Those lessons remain relevant, but AI introduces additional considerations that deserve explicit attention within existing educational frameworks.
Students now need to understand that visual and audio recordings can be fabricated in ways that may be difficult to recognize through casual observation. They should be taught how AI generated content is created, why it can appear convincing, and why a video or voice recording should not automatically be accepted as proof that an event occurred. This is particularly important when the content depicts another student engaging in behaviour that could damage their reputation or result in disciplinary consequences. Educators should also understand that AI detection tools are imperfect and that the absence of obvious visual inconsistencies does not establish authenticity. Verification should involve considering the source, context, corroborating evidence, and circumstances surrounding the material rather than relying exclusively on whether a video appears real.
Schools may also benefit from reviewing their existing codes of conduct and cyberbullying policies to ensure that AI generated impersonation, humiliation, and non-consensual manipulation are adequately addressed. In many circumstances, this may not require the creation of an entirely new policy. Existing expectations surrounding harassment, bullying, respect, and responsible technology use can often be updated to explicitly recognize the misuse of AI generated content. What matters is that students, staff, and families understand that using AI to manufacture humiliating or misleading content about another student may be treated as a serious behavioural concern, even when the material does not involve nudity or sexual content.
We have also addressed the importance of explicitly recognizing harmful AI generated content within school policies. In our article, “A Draft Policy for Schools Specific to Smart Glasses, Wearable Recording Devices and AI Enabled Wearable Technology” (7), we explored the need for clear expectations surrounding recording, privacy, consent, and the misuse of AI enabled technologies within educational environments. Although that article focused primarily on wearable recording devices, the underlying principles are equally relevant to synthetic media created through other applications. Schools should assess the behaviour and resulting harm rather than focusing exclusively on whether the manipulated content contains nudity or sexual imagery. A fabricated video intended to destroy a student’s reputation or encourage classmates to ridicule them may warrant intervention even when the content is entirely non-sexual.
When incidents occur, schools should also distinguish between deliberate malicious behaviour, impulsive poor judgment, and situations where students may not have fully understood the consequences of their actions. These distinctions can help inform proportionate interventions that combine accountability with education. In some circumstances, a restorative approach may be appropriate, provided the targeted student is not pressured to participate and their safety and dignity remain central to the process. In more serious situations involving persistent harassment, threats, significant reputational harm, or repeated targeting, stronger disciplinary measures or law enforcement involvement may be necessary.
Ultimately, we believe schools should be working toward a culture where students understand that another person’s identity is not simply raw material for digital entertainment. The ability to manipulate someone’s likeness through AI should come with an equally strong understanding of the responsibility to respect that person’s dignity.
What Parents Should Do if Their Child Becomes the Target of a Non-Sexual Deepfake
If a parent or caregiver discovers that their child has been targeted through an AI generated image, video, or audio recording, we believe the first priority should be supporting the young person rather than immediately focusing on identifying or punishing the creator. Begin by listening carefully to what happened, how the content was discovered, where it has been shared, and how the child is feeling about the situation. Avoid minimizing the incident simply because the content is artificial or does not involve nudity. At the same time, try not to respond in a way that unintentionally increases the child’s fear or anxiety. A calm, supportive response can help the young person feel that they have someone in their corner who will work with them to address the situation.
Where appropriate, preserve relevant evidence before attempting to have the content removed. This may include screenshots, messages, usernames, dates, links, and information about the accounts or group chats involved. Parents and caregivers should be mindful not to unnecessarily redistribute the harmful material while collecting evidence. If the incident involves classmates or is affecting the school environment, contact the school and provide the relevant information so that staff can assess the circumstances and determine an appropriate response. It may also be useful to report the content through the relevant platform’s harassment, impersonation, privacy, or synthetic media reporting mechanisms, recognizing that the availability and effectiveness of these processes can vary.
If the deepfake includes threats, extortion, persistent harassment, or other potentially criminal behaviour, contacting police may be appropriate. However, not every incident will require law enforcement involvement, and many situations can be addressed through cooperation between the family, school, and platform involved. The response should reflect the seriousness of the behaviour, the extent of the distribution, and the needs of the young person affected.
Perhaps most importantly, parents an caregivers should reinforce that an AI generated depiction does not define who their child is. A fabricated video, manipulated photograph, or synthetic voice recording is not evidence of their character, values, or identity. Young people who become the targets of digital humiliation need to know that they are not expected to manage the situation alone and that there are adults who will support them in restoring their sense of safety and control.
The Bigger Picture: Why AI Literacy Must Include Human Element Literacy
One of the concerns we have repeatedly raised at The White Hatter is that conversations about digital safety often become overly focused on the technology itself. We debate which applications children should be permitted to use, what age they should be allowed to access certain platforms, which parental controls are most effective, and whether legislation can adequately protect them from emerging risks. These are legitimate questions, and technology companies should absolutely be expected to take reasonable steps to reduce foreseeable misuse of their products. However, focusing exclusively on technological restrictions risks overlooking the human behaviours that frequently sit at the centre of digital harm.
This philosophy is consistent with what we explored in our article, “Why Technology Alone Can’t Solve Our Digital Literacy and Internet Safety Challenges” (8). In that article, we argued that many of the challenges young people encounter online are not fundamentally technological problems, but human problems involving relationships, trust, consent, identity, power, and behaviour. Although parental controls, monitoring applications, content restrictions, and platform safeguards can play an important role, they cannot independently teach a child empathy, integrity, respect, or responsible decision making.
The misuse of AI generated deepfakes provides another example of this reality. A student may fully understand how to create a synthetic video while lacking the social and emotional judgment to recognize why using that capability to humiliate a classmate is inappropriate. Similarly, a platform may clearly label a video as AI generated, however that label will not necessarily prevent students from sharing it for the purpose of ridicule. This is why we believe digital literacy education must place greater emphasis on developing the human skills that guide how technology is used, rather than relying exclusively on technological measures intended to control access.
This is also why our approach to digital literacy places such importance on what we call Human Element Literacy, which we believe should form the foundation of responsible technology use (9). Human Element Literacy involves developing the social, emotional, ethical, and critical thinking skills that help young people navigate their increasingly connected lives. It includes understanding how our actions affect others, recognizing the importance of consent and personal boundaries, evaluating information critically, resisting harmful peer pressure, and accepting responsibility for the consequences of our decisions.
When we examine AI generated digital peer aggression through this lens, the central issue becomes much clearer. The problem is not simply that a student has access to an application capable of manipulating another person’s face or voice. The deeper concern is whether that student possesses the empathy, character, judgment, and social awareness necessary to recognize when using that capability would be inappropriate or harmful.
We believe this is where parents, caregivers, and educators can make a meaningful difference. Rather than simply telling young people what they are prohibited from doing with AI, we should be helping them understand why certain behaviours violate another person’s dignity and how their choices can affect the people around them. This approach encourages young people to develop an internal sense of responsibility rather than relying exclusively on external rules, restrictions, or the fear of consequences.
In our opinion, this is particularly important because adults will never be able to anticipate every new application, feature, or trend that emerges within the rapidly evolving AI environment. However, the principles of empathy, respect, consent, integrity, and responsible decision making remain relevant regardless of which technology young people are using.
Artificial intelligence has opened the door to remarkable opportunities for creativity, communication, learning, and self expression. Youth and teens are growing up in a world where they can generate images, videos, music, and other forms of digital content that would have required specialized equipment and professional production resources only a few years ago. We believe there is significant educational and creative value in helping young people understand and responsibly use these tools, and we should not allow legitimate concerns about misuse to overshadow the many positive opportunities that AI can provide.
However, with these opportunities comes an equally important responsibility to recognize how the same technology can be used to target and harm others. The growing accessibility of non-sexual deepfake technology means that digital peer aggression can now involve more than insulting messages, embarrassing photographs, or harmful rumours. A student’s face can be placed into a fabricated video, their voice can be imitated, and their identity can be manipulated in ways that create confusion, embarrassment, social conflict, or reputational harm. None of this requires nudity or explicit sexual content to become a serious concern, and we believe parents, caregivers, and educators need to broaden their understanding of AI related harm accordingly.
In our article, “Guiding Youth to Use Artificial Intelligence Ethically, Thoughtfully, and Creatively” (10), we emphasized that teaching young people how to operate AI applications is only one part of meaningful AI literacy education. Equally important is teaching them how to think critically and ethically about what they are asking these systems to create, why they are creating it, who may be affected, and what consequences could follow. This principle is particularly relevant to non-sexual deepfakes because the technical process of creating an altered image or video may be relatively simple, while the ethical considerations surrounding consent, dignity, reputation, and potential harm can be far more complex.
We believe young people should be encouraged to explore the creative possibilities of artificial intelligence, but that exploration should be accompanied by a clear understanding that another person’s identity is not simply digital material available for unrestricted manipulation. The question should never be limited to whether AI can create something, but should also include whether creating and sharing it is responsible, respectful, and appropriate.
At the same time, we must resist the temptation to characterize every young person experimenting with AI generated content as a potential bully or every face swap as a serious threat. Most youth and teens who explore these tools are doing so out of curiosity, creativity, humour, or a desire to participate in the digital culture of their peers. Our responsibility as adults is to help them recognize the difference between creative expression and the exploitation of another person’s identity for entertainment, social status, or humiliation.
We should also remember that artificial intelligence did not create the human desire to embarrass, exclude, or intimidate others. Those behaviours existed long before social media, smartphones, or generative AI. What technology has done is provide new mechanisms through which those behaviours can be expressed, amplified, and distributed. This is why we believe the conversation must extend beyond the applications themselves and address the social and ethical decisions being made by the people using them.
Rather than focusing exclusively on whether a particular AI application should be restricted or banned, we should also be asking whether we are adequately preparing young people to make responsible decisions when using increasingly powerful digital tools.
Are we teaching them to seek consent before manipulating another person’s likeness?
Are we helping them understand that humour should not come at the expense of someone else’s dignity?
Are we encouraging them to question the authenticity of what they see and hear online?
Are we giving them the confidence to challenge harmful behaviour when their peers use AI to target someone else?
These are the conversations that need to be taking place in homes and classrooms today, not after a young person has already become the target of a humiliating deepfake.
At The White Hatter, we believe one of the greatest challenges presented by AI generated deepfakes is not simply that technology can now make people appear to do or say things they never did. It is that technology can give someone the power to manipulate another person’s digital identity without their knowledge or consent, sometimes for no other reason than to entertain an audience at that person’s expense. Teaching young people to respect that boundary is not simply an AI literacy issue. It is a matter of character, empathy, dignity, and responsible digital citizenship.
As we have often stated, our goal should not be to raise children who are afraid of technology, but rather to raise children who possess the knowledge, skills, confidence, and character to use it responsibly. In a world where artificial intelligence can increasingly manipulate what we see, hear, and believe, the human element of how we treat one another becomes more important, not less. The technology may be artificial, but the dignity of the person being targeted, and the harm they may experience, are very real.
Digital Food For Thought
The White Hatter
Facts Not Fear, Facts Not Emotions, Enlighten Not Frighten, Know Tech Not No Tech.
References
