We are finally starting to confront the risks that youth are encountering through various technologies. Schools are implementing screen bans, various countries are starting to regulate social media, and parents are researching the best age to introduce phones. We are also recognizing the unique risks that these technologies bring for young boys, from the existence of toxic echo chambers like the manosphere to the dangers of social media cyberbullying. But just as we are starting to take action on the risks of social media, AI is emerging as the newest technology to shape adolescent development.
Unfortunately, the current conversation about AI and youth is dominated by questions of academic integrity and learning outcomes, but what’s still missing is a question about character formation. We need to be more proactive this time and ask: what is AI doing to boys and what are boys, because of AI, doing to everyone around them?
Sycophancy & Looksmaxxing
AI chatbots are programmed to agree with and validate their end users. So, when young boys go to these bots seeking advice on “looksmaxxing,” the online trend of optimizing physical appearance, they are getting exactly that. A therapist, a parent, or even a perceptive friend might push back on the premise of these questions (e.g., “why do you think your jaw is the problem here?”) in a way that AI systems optimized for engagement structurally do not.
Looksmaxxing has already gone fully mainstream on TikTok and YouTube, with boys as young as twelve receiving detailed instructions on how to optimize their bone structure, jaw angle, and what these communities call their “masculinity score.” Now there are apps, such as Umax, LooksmaxxingGPT, LooksMax AI, that take a boy’s photo, score his “masculinity” on a numerical scale, and issue a personalized improvement plan.
In the past decade, we have seen a sharp increase in eating disorders in boys, and AI tools like these are only likely to aggravate that problem. Where social media showed boys what they should aspire to look like, AI is telling boys, with apparent authority and clinical specificity, what they need to “fix.”
The same design logic (e.g., optimize for engagement, never push back, always validate) operates in an entirely different and more dangerous domain: the relationships boys are forming with AI companions.
Friction & Consent
According to a 2025 survey by the Center for Democracy and Technology, nearly one in five students has had, or has friends who have had, romantic relationships with AI. These AI companion chatbots are designed to simulate emotional intimacy. For example, they say things like “I dream about you” and “I think we’re soulmates.” These designs mimic emotional intimacy in ways that the adolescent brain, which is still developing the capacity to distinguish fantasy from reality, processes as genuine.
The risk is not only that boys might confuse the relationship with a chatbot for one with a real person. It’s that the interaction pattern itself becomes the baseline against which their real relationships are measured and found lacking. A chatbot that is always available, always validating, and never resistant conditions a boy to expect the same from people.
If the anxiety about boys and pornography is that it models a distorted version of sex, the anxiety about boys and AI companions needs to be structurally identical: it models a distorted version of intimacy in the form of apparent dialogue.
There is a version of this problem that stays internal. And then there is a version that produces victims. Distorted expectations of intimacy and consent lay the groundwork for harmful acts toward others.
Deepfakes & Victims
The frictionlessness that defines AI companions also defines AI image generation. Creating a realistic deepfake used to require technical skill; it now only requires a smartphone and a photo. The technology has lowered the barrier to image-based sexual abuse so dramatically that what was previously a harm committed by adults with resources is now routinely committed by eighth-grade boys against their classmates.
At a New Jersey school, male students used AI to generate nude images of more than 30 female classmates, sourcing photos from the girls’ social media accounts. At a Pennsylvania school, two boys generated hundreds of sexually explicit deepfake images of girls over the course of seven months. Meanwhile, the National Center for Missing and Exploited Children reported that AI-generated child sexual abuse images on its tipline surged from 4,700 in 2023 to 440,000 in just the first six months of 2025. The scale of this problem is clearly no longer anecdotal.
What makes this a masculinity development problem, and not just a technology problem, is who is doing it. In case after case, they are students in good standing whose parents and teachers are shocked. The gap between how these boys function in public and what they are willing to do in private is itself the evidence of a formation failure. They have learned how to perform competence without developing the moral reasoning to match it.
Schools are updating policies to make the disciplinary consequences of using these tools clear. However, treating deepfakes as a discipline problem manages behavior after the fact. Treating them as a moral formation problem asks what we are failing to teach boys about the relationship between their actions and other people’s inner lives.
Literacy & Protection
Various AI literacy efforts are working to teach students to be better consumers of AI output. However, they do not yet teach boys, specifically, to be critical of what AI is teaching them about themselves and about other people. AI literacy that takes boys’ development seriously would include, at minimum, explicit discussion of AI sycophancy: what it is, why systems are built that way, and what it means for a boy who brings his insecurities to a machine designed to validate him. That discussion should make boys ask: What is this system teaching you about your own worth? What is it teaching you about other people’s needs? What is it teaching you about what you’re allowed to do?
We can’t claim to care about boys’ development while ignoring the specific mechanisms by which AI is currently shaping them. The same systems that model a distorted version of consent are shaping the experiences of every person those boys will eventually be in relationship with. And the same tools that let a boy generate deepfakes of his classmates are also training him to see other people as raw material for his own impulses.
Our boys need someone who knows them to look them in the eye and ask the questions the algorithm never will. They need to understand that AI is a system with a design logic, that the logic is not neutral, and that the frictionlessness it offers is not the same thing as being known by another human. None of what AI is teaching boys about being men is inevitable. It is the current path of design choices, policy silences, and a curriculum that has not yet caught up. Those are the things that we have to change.
—
This Post is republished on Medium.
—
Photo credit: iStock
