Interview: How Does Violent Pornography Affect Young People?

Violent pornography is easily accessible on the internet, and some children encounter it before they even get to middle school. How does this influence their expectations when it comes to physical intimacy? How does exposure to such content influence students on college campuses, where sexual assault is common and often unreported?

Young people may be particularly vulnerable, says Melissa de Roos, a forensic psychologist at Erasmus University Rotterdam in the Netherlands. That’s because they lack the life experience to be able to tell when they’re watching an extreme or distorted version of reality. Researchers who want to better understand this, however, face a significant methodological challenge: how to design studies that mimic the real world but don’t cause any harm to the people who have volunteered to participate in the research.

“I do think we have a responsibility as society to equip children, young people, so that they have the tools to navigate the online spaces that they are in,” said de Roos.

Undark recently spoke with de Roos over Zoom to discuss where all of this is going, particularly given the rapid proliferation of AI videos of child sexual abuse. Our interview has been edited for length and clarity.

Undark: How did you get started in your field, and what motivated you to study the links between sexual violence and online behavior?

Melissa de Roos: I started a long, long time ago just purely looking at sexual violence, and then perhaps more specifically, focused on child sexual abuse and child sexual abuse material, which more colloquially is called “child pornography,” but we don’t really like the term.

UD: Can you talk about why you don’t like the term? I see this acronym all the time: CSAM.

MdR: “Pornography” suggests it’s for entertainment. It’s — I don’t like the word, but — titillating, stimulating. It’s acceptable. Whereas, really, what we are talking about are images, or videos, or just content depicting sexual abuse of children. I think to call it “child pornography” sanitizes what it actually is.

I became quite interested in how we’re seeing such a rise in CSAM offending, but the explanation clearly can’t be that suddenly we have so many more people with a sexual interest in children. There must be something else going on. What I’m particularly interested in is not so much the people who sit down with their laptop and search for this kind of content, but people who perhaps are continuously looking for something more extreme, something new, something taboo that they haven’t seen before. For some people, that can end up with them looking at various types of illegal content, including child sexual abuse material.

That fed back to violent pornography. There’s so much. It has become so mainstream.

Given that it is so easy now for even very young children to access this stuff, what is that doing to the world, to the people?

UD: What is that doing? It’s hard to study, right?

MdR: Yeah, it is. Because if you want to ask underage children about pornography and sex, that is obviously ethically a bit tricky. For that reason, up to this point, I’ve focused on people over 18, but we are definitely seeing that that is too late because you’ve got kids as young as 10, 11, continuously watching quite extreme pornography.

UD: That’s sad.

MdR: It is sad. There are a lot of good things that have come with the internet also. Certainly for kids perhaps who are exploring their sexuality, their gender identity, anything related to that. There’s now a place online where kids can go and say, “Hey, is this normal?” They can talk to peers and exchange experiences. In that sense, it has become, I think, quite a valuable resource for sex education.

But this is the downside of that. They are also coming across things that because they are so young, because they probably don’t have real world experiences, this is what they’re basing their ideas on.

We are seeing reports coming out of England where they’ve got primary school pupils asking in sex education about choking and that sort of thing. It makes sense: They are young, they don’t have other experiences to balance it out, so they’re much more vulnerable in that sense. And if this is what they see online, then they probably, in the absence of something to counterbalance that idea, they’ll adopt that and think this is just normal, this is how it’s supposed to go, which, of course, creates quite a lot of problems later in life.

UD: In a 2025 paper on violent pornography and sexual aggression, you explore how viewing violent pornography might shape people’s expectations in real life. Why is it important to look at perceived realism and peer beliefs when studying these links?

MdR: Violent pornography, in and of itself, does not typically or necessarily lead to real-world sexual violence. You tend to need to have other ingredients present as well.

Now, if you are watching something and you think, “Wow, that is insane, that is so not realistic,” you are going to be less likely to adopt that into your blueprint of how a relationship works or a sexual encounter works. If you think “This is fake, these people are being paid,” etcetera, it’s less likely to have quite such a big effect. Or if, for example, you have had sex education in school and these issues have been discussed: How does the porn industry work? Porn literacy, media literacy programs, that kind of thing. Then when people think, “Okay, this just isn’t realistic,” the effect isn’t as great.

Peer beliefs are another ingredient. If you are watching such violent content, and so are your friends and you are talking about this, then your friendly discussions with your peers — about relationships, about sex, about girls, about women — they are going to be influenced and shaped by that. A peer effect where everybody kind of has those same ideas and nobody challenges them, that is very powerful. Because teenagers, they tend to want to fit in with the people they hang around with. That exacerbates the potentially negative effect of watching that kind of content.

UD: Do you have thoughts on when is the best time to start teaching porn literacy? I think we have a situation — I’m just speaking as a parent for a moment — where there’s a sense that kids are too little to have this conversation, yet they are encountering violent pornography online. So it’s also harmful to not address it.

MdR: I think it’s on a child-by-child basis. But if your child has a phone and uses it unsupervised, then it’s a conversation, an age-appropriate conversation. I don’t think you should sit them down and say, “Well, ta-da!’’ This is not how it should go. But you need [to offer] context: “This is not necessarily healthy. These people are being paid. You might come across some things. Come talk to me if you have any questions.”

If they’re old enough to have a phone without supervision, then I’d say they’re old enough to have those kinds of conversations. They don’t need to be explicit, or extreme, or scare them, but keeping an open door, I suppose.

UD: What are some of the biggest methodological challenges to studying sexual violence and online behavior?

MdR: We’re relying quite a bit on self-report. In that 2025 study, I am asking people, “Have you ever done the any of these things?” [The possible answers] range from “touching someone when they didn’t want me to” to “forcible violent rape” and everything in between — there’s “coercion” in there. We are hoping that people will be open on an anonymous survey.

For a study I’m starting in two weeks, I’m trying to figure out what is actually going on when people interact in a sexual situation, which becomes an ethical issue: How do you do that ethically in a lab environment? I received a grant to create AI-generated stimuli [with a collaborator] who can make hyperrealistic video vignettes, not explicit, but of an ambiguous sexual situation, where there is some coercion, whether that is verbal or physical.

We’re going to have people look at it, and rate what they think is going on in that situation at that point. What we’re hopefully going to be able to do with that is figure out, what are people paying attention to when they determine whether the person they are sexually engaged with is still having a good time? What is their perception of consent? When does it shift? How clear or obvious does something need to be?

We can hopefully take a first step into looking under the hood, if you will. What is actually going on? Because if it is a case of, they’re not paying attention, or they’re looking for the wrong things — someone may say, “Yes, this is fine” but actually non-verbally be completely not fine — then that is something that you could incorporate in, for example, sexual safety programs on campuses, which often have huge problems with sexual violence.

So, I think the ecological validity — mimicking real world situations, without doing harm to participants — is the biggest challenge here.

UD: Where do you see this research field going in the future?

MdR: That’s an exciting question. In this field, we tend to focus a lot on the negative aspects of AI. I do think we should also perhaps think about ways in which we can actually use it for good. I know, for example, various police forces are using AI to screen CSAM and trying to figure out if it is real or not, which takes a huge burden off the police.

UD: By real, you mean actual kids, not generated by AI?

MdR: Yeah. Those kinds of developments. I do think with care, we should try to use this for good, instead of just letting it drag us into dark places.