Are porn algorithms feeding a generation of paedophiles – or creating one? | Hacker News Reader

Are porn algorithms feeding a generation of paedophiles – or creating one?

theguardian.com

43 pointssandebert63 comments

Open articleView on HN

They are doing neither, and the article itself undermines its premise later when the "serious research" is being quoted. As it cites, most offender get exposed during early teens, back when there were no porn algorithms. However, the rate of child sexual abuse has steadily decreased since the 1990s, down by almost 50%. If the hypothesis of the article was correct then the rate of child sexual abuse should increase, not decrease.

Other research studies, have looked at the rate of pedophilia in the general population, and the conclusions seems to be a fairly static rate of around 2% (like "Prevalence, situation, and perspectives of treatment" from 2020). However since the rate of child sexual abuse has decreased, such studies mostly rely on doing survey and those has for obvious reasons a major problem of sampling and accuracy.

A common theme in the research, just from a very quick look, is that the researcher themselves cite that the is a huge lack of research in this area. They can see the trend, and see the numbers, but there is little to real understanding to what is behind the numbers. We currently has as good chance to blame the Catholic Church as to blame pornhub, and articles like this one are not helping the slightest. At best it just spreading fear in order to generate engagement, and at worst they are abusing their readers by pushing propaganda in order to sway popular opinion about oppressive laws.

> the rate of child sexual abuse has steadily decreased since the 1990s, down by almost 50%.

That's kind of remarkable if true, because the openness of both individuals and institutions to taking reports of it seriously has increased. Before all our modern awareness of stuff like "consent" a lot of reports were dismissed, especially if the pedophile was someone trusted in the community and especially if they came from women. Boys often didn't report at all out of shame.

Usually when there's an increase in awareness you see a phantom increase due to reporting, as has happened with things like rates of homosexuality or with diagnostic categories like autism and ADD/ADHD.

Of course maybe there's less of it today because abusers are getting caught more often. That's possible too.

Edit: this wouldn't surprise me at all given that there appears to be zero relationship between public panics about something and its actual incidence. Street crime, for example, was orders of magnitude worse in the early 1990s vs. today but people were less panicked about it. Public panics are media driven.

I'd certainly also like to see a source for the basic claim. It would be wonderful news if true.

Why don't you Google it? This is fundamental basic knowledge about CSA incidence.

If it's so trivial, you should find the claim no harder to substantiate than to insist upon as here.

I encourage the attempt. Seriously. Go show me how easy reliable data is to find on this. Prove me wrong and you'll have earned a friend for life, or at least a dinner at the nice place down the street the next time you're in town.

OK... but there really are so many results upon literally just searching it. You immediately get multiple prominent results. You'd think if people actually cared about substantiating this, they could do one (1) Google search to find out. But here you go, courtesy of lazyweb.

https://files.calio.org/BIBS/declining-rates-bib4.pdf

https://www.stopitnow.org/sites/default/files//prevalence-ra...

https://www.unh.edu/ccrc/resource/decline-sexual-abuse-cases

(Of course the answer is that people don't care. They'd be happy to ignore the lack of sources if the conclusion being drawn agreed to them, but can't do a single fucking Google search if it doesn't.)

I'd be interested to know what search query you actually used. It's also been a while since I compared DDG and Google result quality. Maybe I'm getting poorer results than in the past. Queries on the theme of "child sexual abuse statistics US by year" don't get me anything very useful from DDG, much more advocacy around headline prevalence figures than anything relevant to incidence. (Statista does reflect a drop, but I didn't want to make an account to see their sources. Fair, my fault, I can't evaluate that well.)

The bibliography is interesting and useful but not immediately evaluable, and seems mainly Canada-focused. The Stop It Now/D2L data is very clear up front that it relates to prevalence and not incidence, a distinction that matters for reasons to which we'll return in a moment. The UNH data substantiates some level of decrease since the early 90s, as does every other source, but disagrees with itself on magnitude in different places and comes out of one researcher's shop, which does not invalidate it but does mean I'd like to see someone else produce something similar before I weight their novel results too heavily.

I guess maybe we have differing definitions of "reliable," but I think you and I have clashed on this subject before. I don't blame anyone for assuming that a victim of such abuse would tend to overestimate its prevalence, and I try to watch out for the tendency in myself, but what's all the anger about?

The difference between prevalence and incidence matters because the former tells us how many people have been abused as children, and the latter tells us how frequently children are abused over time, which is the thing we're actually talking about. How we count multiple instances of abuse against the same person confounds everything, and intuitively I would expect prevalence by victim (ie for a given victim how frequently they themselves were abused) to follow something like a power-law distribution, but in general if we just take the headline number of "X total kids" it's hard to see how that gets us to a nuanced understanding of incidence rates that lets us talk knowledgeably about how and to what extent the shape of this problem is changing over time.

So when you said:

> I'd certainly also like to see a source for the basic claim. It would be wonderful news if true.

You actually already knew about all of these sources and have debated this before, and decided now to start talking about this bit of nuance?

I'm not going to play. Sorry.

I'm familiar with the subject through long study. The sources you identified were not previously known to me but I feel confident evaluating them as I have based on prior knowledge and review. Do you mean to argue whose web search skills are the best - a point I already spent a paragraph conceding - or do you mean to argue the subject actually under discussion?

I still don't understand where the anger is coming from. Why does it piss you off that I want to argue child sexual abuse is more common and more of a problem than it's broadly acknowledged to be? Are you assuming I mean to use that claim as an argument against porn or something? I really can't figure out what's gotten so far up your nose on this, and I wish I could because it's really making it impossible for us to have a conversation.

Last I recall, we spoke cordially enough last year, if not generally in agreement, on the subject of AI capable of CSAM generation, and how it might be regulated. That was coincidentally a very strange day for me, granted, but again, I struggle to see what's changed since to earn this sort of acrimony.

With whom do you understand yourself to be needing to have this fight right now?

> I still don't understand where the anger is coming from. Why does it piss you off that I want to argue child sexual abuse is more common and more of a problem than it's broadly acknowledged to be? Are you assuming I mean to use that claim as an argument against porn or something? I really can't figure out what's gotten so far up your nose on this, and I wish I could because it's really making it impossible for us to have a conversation.

It's pretty simple honestly, any online discussion about pornography has turned into a complete and total clown show and it has become increasingly obvious that people who are arguing are doing so in bad faith.

I don't know what previous conversations we've had, as I spend far too much time on this damned website and have probably made tens of thousands of comments, but I kind of had a bad feeling the moment I saw your first comment because I had a pretty good feeling you knew what sources people were referring to and decided to play it as if you didn't, for one reason or another. I can see strategically why one would do this when the data seems to contradict the conclusion you have drawn so unilaterally, but at the same time, it's an annoying strategy. It's like laying a trap.

I really don't like this sort of discussion thread:

> Oh, it would be great news if incidence of CSA has gone down. Care to link to a source?

> Sure, here: [d2l link]

> Ah, but, you see, that talks about prevalence, not incidence. [...]

Because you knew what direction it was going to go all along. Playing coy here feels like theatrics.

If I took your position, I'd probably go more along the lines of:

> I am aware of the abundance of studies that discuss the prevelance of CSA, but [...]

(And I'm sure people would probably downvote that, but also, I get downvoted any time I respond in one of these types of threads, so it wouldn't make any difference to me.)

Anyway, the truth is that when I saw your first post, I didn't know what type of trap was being set, but my expectations were pretty low. The actual nuance you brought up is more interesting than that, but still not something I really care to debate.

I'm completely lost.

You raise sources I've never seen, I evaluate them as fairly as I can, you don't like the result so I'm arguing in bad faith.

I mentioned incidence, one source raised to try to counter not only relates to prevalence rather than incidence but explains why this matters, but when I point out that a source you chose itself disclaims support for how you want to use it, I'm arguing in bad faith.

I still don't even know what kind of pernicious dishonesty I'm meant to be advancing here, but sure, if you like then apparently I'm setting some kind of trap, in service of good God, I still don't know what.

I don't know who you imagine yourself to be at war with here, but whoever you're talking to clearly isn't me, so...I don't know, good luck with the rest of your day, I guess. You're not wrong we both have better ways to spend our time, at least.

I can explain a lot of the disconnect: I simply don't believe you.

For one thing, because the decrease in CSA over time is so well-known that I am pretty sure one of the PDFs I have saved refers to it as having been "common knowledge" in the 2000s. I have a hard time understanding how someone could apparently have studied this topic in detail and not come across this.

For another, I posted 3 sources and you 'evaluated' them within 15 minutes of me posting them, and of course found issues with all of them.

Yes, I can see that you are replying cordially, but that doesn't necessarily mean you're being forthcoming, either. I've grown extremely tired of this type of discourse.

Currently it's 2025 - in the early 2000s when most of the sources you linked come from there wasn't even Facebook let alone pornhub. The ones reporting on more recent times seen to have consensus that factors in reporting, especially covid, make it difficult to come to any conclusion. So a fair point of discussion that seems to be a far cry from "common knowledge". I don't know what you're aiming for but it's hard to take it in good faith really.

My basic search turned up:

https://www.unh.edu/ccrc/trends-child-victimization

Scroll down to the "Updated Trends" link and click it for more up-to-date info.

In the latest one (looks like it covers 2023, released last month), child sexual abuse is down 67% since 1990. (Also, physical abuse is down 63% and neglect down 27%.) Child sex abuse down 63% from the '90s.

Since you turned up one of the same sources as my other interlocutor, I'll refer you to my reply there at https://news.ycombinator.com/item?id=43593671

I would guess that pedos today are far less likely to commit sexual abuse given that the chances of being caught are high and the repercussions are extreme. In the past this was much less the case.

Perhaps? With no evidence, it strikes me as a crime that is not weighed in ROI terms; it’s people with a sick compulsion, and I have always believed that kind of thing isn’t affected by the severity of possible punishment. I’d be interested in data for sure.

With no evidence the proper assumption would be no assumption.

But evidence was presented child sexual abuse declined since 1990. The question was how to explain it.

Punishment severity doesn't much reduce child sexual abuse probably. But not because it's a sick compulsion. Punishment severity doesn't much reduce most crimes. Financial crimes included.

They mentioned chances of being caught also. Punishment likelihood is an effective deterrent generally.

Pornography could be a factor. Many studies found rape decreased in countries when pornography became more available.

You're way off but I don't feel like taking the opposite side of this argument. If you want to research this there's plenty of data and information available.

Indeed, people like to label anyone who is attracted to people under the age of consent as "pedophiles" but the definition is as you point out.

https://en.m.wikipedia.org/wiki/Chronophilia

There is a categorical difference with your examples for "phantom increase due to reporting". On one side it's something done to you/someone, in particular a serious crime, on the other side it's something about you/someone, in particular not a crime.

This is important, because the risk of getting caught, not necessarily severity of punishment, is the main factor in crime deterrence, as far as I know. I presume, increased vigilance and ease/encouragement of reporting has a significant deterring effect on potential perpetrators. As with all sexual abuse, the vast majority of cases happens in settings where the victim knows/trusts the abuser beforehand, therefore conviction entirely depends on the victim's willingness, prospect and ability to name/accuse the abuser. Coming with the reporting, there is probably also greater awareness and knowledge about the manipulation tactics and situations enabling abuse, helping guardians, friends and authorities to identify concerning dynamics and encounters.

However, I presume, the increased ease of reporting and awareness also indirectly prevents further abuse by subsequently encouraging victims to seek help and work with their trauma. I've heard a significant share of perpetrators have been victims of abuse themselves in the past, so seeking therapy after being victimized may also break the cycle of violence.

We should be careful not to equivocate between pedophilia which in the technical jargon of research studies is a sexual attraction to pre-pubertal children and child sexual abuse which loosely is the act of sexually assaulting someone below the legal age of consent. They are two different things. An abuser might not be a pedophile in the technical sense, and a pedophile might not be an abuser.

So rates of pedophilic attractions and use of CSAM could be rising while rates of child sexual abuse is falling. I don't know if that's the case, but we shouldn't confuse the two things.

It's one thing when people actually seek it out. But the idea that someone may get charged for a video they didn't realise contains underage people is chilling. I know one tiny woman who in her 30s still could pass for a highschooler and will probably be carded for many years to come. On the other hand I went to high school with a couple of people who could pass for 25+ year olds easily. Without personally checking someone's id, you just can't know for sure. So even if you'd want to avoid illegal material - how sure are you you're safe?

All commercially produced porn is fastidious about record keeping. Even to the extent that someone underage used a convincing fake ID, relying on that infra would be a pretty strong defense. Truly amateur, unknown source, “teen” content is scary as hell.

> Even to the extent that someone underage used a convincing fake ID, relying on that infra would be a pretty strong defense.

Fake ID is irrelevant in many jurisdictions.

It’s moot because the consumer never sees these IDs. But you can’t with a straight face say there is as much potential liability for someone watching “teen” porn from a legit company that includes 2257 statements as there is from watching torrented, unsourced, amateur “teen” content. It’s a huge difference in intent.

Well I don't know about the "porn algorithms", people are free to choose what they want. So if they continuously choose from the "teen" category, I'm guessing some sort of deviation must be at play.

Like if you check the teen category on https://www.pornpics.com/ , many of the girls there seem awfully young with some, practically kids.

For me the feeling is too uncomfortable to watch those. But I did notice a curious phenomenon, now that I'm approaching 50s. There were all kind of movies I watched as a kid, teenager or young adult and I completely ignored the beauty of the actresses in them, they were at best associated with "mom" if not "granny". But now, that their age in the movie is 10-20 years below my current age, suddenly I discover that they're pretty good looking. Like Elaine from Seinfeld for instance :)

I heard somewhere this saying and I'm afraid it's true: "When a man is in his 20s, he finds beautiful maybe 5-10% of the women of his age. When they're in their 40s, they find about half the 20-somethings beautiful. When they reach 80s, they find all of the young women beautiful." :)

Personal anecdata: while approaching 40, I wouldn't say anywhere near half of 20 somethings are beautiful.

Not even porn algos, regular algos provide a feast for those people. Tiktok, previously musically, youtube, it is all filled with questionable underage content. The worst part though is that if you are into such things, it recommends more to you, and other similar users history and activity gets recommended to you, thus distributing the content. It is no different from how these algos learn you like mountain biking or gardening.

I hate how when an article about how science has discredited the concept of porn addiction wound up getting bounced off the HN frontpage within minutes of appearing on it, then flagged and deaded:

https://news.ycombinator.com/item?id=43587448

And yet, this absurd sensationalist dreck will probably stay up here for longer.

Could you say more about this? This link isn't that helpful since the main content was removed.

The link went to here: https://www.psychologytoday.com/us/blog/women-who-stray/2018...

Any addiction is bad, including porn. Why normalise it? Addictions can be greatly advanced by an algorithm, you can't deny that either. Are porn sites trying to advance porn addiction in their users? Well, the number of views is their money, so it's reasonable to assume they are.

I could agree with only one thing, that porn addiction is not the worst one out there, but blamed the most. For some reason, it's ok for mobile games to make gambling addicts from kids, but no one seems to care about that.

I don't really believe porn addiction ever really was normalized. The layperson definition of "addiction" is just something that someone is habitually occupied with. The medical definition of "addiction" is more like something that you have a serious dependence to, to the point where it impacts your life and it would cause distress to be without it.

And don't get me wrong, I am not saying that there are not people with a sex or porn addiction, but I think it is woefully uncommon. I'm pretty sure if you take the average person that the Internet classifies as a "gooner" or "porn addict" and took away all of their porn, they would be completely fine. It's the same for those stupid mobile games, too. They get the vast majority of their money from a small minority of people, and even then, not all of those whales are addicted either, some of them are just well off enough and choose to spend their money on frivolous things, no different than people with thousands of vinyl figures (and although I think they have a bad taste in figures, I doubt most of them have a serious problem.)

What I do think was normalized in the late 90s and early 2000s was extrarordinary crudeness and overt sexuality, but I don't think there was really a point at which people were normalizing getting psychologically dependent on pornography. It'd be kind of hard to normalize that since it's more about the specific person than it is the thing they're dependent on.

This also does not mean that social media or algorithms aren't fucking with our brains in other ways. And I'm going to be honest, I have absolutely no idea what the porn algorithms are or what they're doing. I don't use Pornhub or sites like Pornhub (yes, I know, "sure you don't", but I'm serious.) I do use YouTube constantly, and I don't think I'm addicted to YouTube. I'm not addicted to Technology Connections or Practical Engineering. I'm sure the YouTube recommendation algorithm is fucking with my brain, but I wouldn't call anything it's doing an addiction. I'm also sure this is true for a majority of TikTok users, even though it's a ton more... well, what us laypeople would call "addicting".

I think the argument is that even very disordered and disregulated consumption of porn doesn't resemble what we typically define to be an "addiction". There's a meaningful difference of degree and of kind between a bad habit that's difficult to shake, even a deeply entrenched one, and medical/psychological addiction, and researchers have repeatedly concluded that porn consumption never reaches the standard of the second.

I'm going with "no".

Betteridge's Law is strong with this one. The lede is buried at after all the sob stories about guys who discovered they were pedophiles because recommendation algorithms are really good as showing you what you want to see. This is all just pretext to the UK government's massive privacy overreach and censorship.

> Along with other countries, the UK is bringing in tighter controls on the porn industry. Porn sites will be required to have age verification processes in place by July to stop children viewing their content, a process that will be monitored by Ofcom under the new Online Safety Act (OSA). But the OSA does not ban or control certain common practices in porn, such as strangulation, sexual fetishisation of incest or the abuse of children acted by adults. The government argues depiction of sexual strangulation is banned but a recent in-depth review of porn by Baroness Bertin, a Conservative peer, warned it is not, in reality, covered by case law. She called for tighter regulation of “legal but harmful pornography like choking, violent and degrading acts … even content that could encourage child sexual abuse”. While headlines picked out the call to ban strangulation, Bertin also demanded clearer rules around “incest, step-incest and ‘teen’ porn”.

I think the people pushing these moral purity bills need a wake-up call about what women find hot. Because this worldview makes no sense once you know that these "violent and degrading acts" are extremely common fetishes among women who play the part of the "victim." It's not men who are reading gothic horror where incest one of the tamer depictions. The problem in the real world is that, despite the public perception otherwise, guys are generally deeply uncomfortable playing out the other side of these kinds of fantasies.

> Because this worldview makes no sense once you know that these "violent and degrading acts" are extremely common fetishes among women who play the part of the "victim."

I’d like to see some evidence for that because it sounds like a projection to me.

Here is some data: https://aella.substack.com/p/fetish-tabooness-and-popularity...

I haven't looked at it closely, so I have no idea whether it supports or opposes the hypothesis. Just wanted to give you some data.

"being submissive, sexual frustration, light bondage, spanking, choking"

All clustered at the top in the feminine-preferred category.

The author of that post is surveying their own audience. You can take the results with a pinch of salt.

> You can take the results with a pinch of salt.

That's a sub-genre I have yet to hear about, do tell more.

Seriously, though, the 'porn addict turned sexual predator' questionnaire in TFA was from anonymous people on the dark webs. Are they also salt-fetishists?

Find your local BDSM community, and you'll have it. Maybe. If you're a good boy.

That doesn’t make it “extremely common”.

“Extremely” is subjective. But research[1] says almost 50% of the population has engaged in BDSM, another 13% fantasizes about it, and 12% engages in it regularly.

I’d call that extremely common, far more common than skateboarding or sports cars, but YMMV may vary in definition of “extremely”.

1. https://academic.oup.com/jsm/article-abstract/14/9/1152/6973...

A third of British women have suffered unwanted choking in the bedroom:

https://metro.co.uk/2019/11/29/third-british-women-suffered-...

Sure? And a shockingly high percentage of women have been raped. How does that impact how common it is for adults to consensually play rough?

I can't even count how many movies I've seen where this is a comedic plot point -- the conservative soccer mom likes to get a little freaky and someone walks in.

+1 with the other reply to your comment.

I don't as an anecdote disagree with some of what you said. There are many things happening in private that you don't hear about in public. Friends aren't telling friends things. And I think it's important that people feel more comfortable/open/vulnerable with people they know rather than projecting what they want people to think.

But I'm a guy and don't claim that I know what women want! Perhaps you're a woman and thus your reply is rooted in lots of experience/evidence?

This strikes me as invidious, an effort to silence. Perhaps I misjudge, so I'll respond in good faith: your prior interlocutor attributes to women what is in fact a common trait of people who've previously been abused.

The misattribution is common; women more often come in for sexual abuse on the one hand, and on the other benefit from much more space in the social narrative to exist as survivors thereof. Men who have been similarly abused tend to respond differently, for that and other reasons.

I'll respond in good faith too: note that I said "anecdote" - maybe I should have said something like "exclusively personal experience"

I thought the OP (in comments) was quite authoritative and confident but without much evidence/support for making such a statement.

Don't you think that's rather bold? - to call someone out on an unsupported assertion, in the very same breath as you say "and no it isn't, probably!"

Yes, I'm a woman posting about my own experience as well as the experiences of my female friends who I could in no way describe as skewing particularly sexually deviant. They all pretty much consider themselves "vanilla" and yet this article would have you believe that they're deeply depraved for wanting to be "taken," choked, dominated, or receive some light degradation.

These things are so normal that they're not even considered to separate from sex, they're just what you happen to be into— on the level of preferring specific positions.

Interesting to hear this. And thank you for that open view. And, yeah, my personal experience lines up somewhat with what you said as well as some stories I hear about other folks. Lot of it sounded like people finding - among the many possibilities - ways to change things and have fun and experiment.

> I think the people pushing these moral purity bills need a wake-up call about what women find hot. Because this worldview makes no sense once you know that these "violent and degrading acts" are extremely common fetishes among women who play the part of the "victim."

Agreed, every single woman I’ve slept with has expressed a desire to either be restrained, submissive, choked or spanked, and a few have expressed desires for nonconsensual sex fantasies. It’s exceedingly common.

all([]) == True

I’m not saying 100% of women are into it, just that 100% of the partners I’ve selected have been into it. I choose partners that are positive and open about sex so your experiences may be vastly different. I’ve never met someone on a ‘kink’ website, just your regular methods of finding a partner.

he's right... it's very, very common. and along multiple generations (well at least current 20 to 50s) and multiple cultures.

I'm not saying you're implying this, but I need to be the boring hall monitor to point out that while consensual submissive fantasies are extremely common there is none -- zero -- overlap with actual non-consensual sexual assault which I'd venture to say nobody (to within a rounding error of zero) finds appealing.

It's kind of like the difference between watching a scary movie or visiting a haunted house attraction vs. actually being chased by a brutal crazed murderer.

I don't disagree which is why playing out a rape fantasy is generally called "consenting non-consent." There are rules, boundaries, safe words, lots of aftercare, just generally good communication all around is a necessary condition.

Haunted House is actually such a good way of explaining it. I'm stealing that!

Reply on news.ycombinator.com