Yours Lawfully Podcast
Yours Lawfully Podcast
Fighting the beast of image-based sexual abuse
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
"Platforms have an obligation to take things down... but the problem is that the thing spreads so fast that the victim would constantly have to issue these notifications to take stuff down. Then there are some platforms that would just refuse."
Every day, intimate images are shared without consent - stolen, or increasingly fabricated with AI. Image-based sexual abuse (IBSA) is now a major driver of sexual offending, and the law is only just catching up.
In this bonus episode of Yours Lawfully, we're joined by Professor Julia Hörnle, Professor of Internet Law at Queen Mary University of London and Technology Law Institute Director. Drawing on her two-part article series "Fighting the Beast," we ask why criminal law alone can't stop IBSA, where platform regulation is falling short, and whether Ofcom is being ambitious enough.
Read Julia's articles:
· Fighting the beast of image-based sexual abuse, Part 1: https://academic.oup.com/ijlit/article-abstract/doi/10.1093/ijlit/eaag006/8673064
· Fighting the beast of image-based sexual abuse, Part 2: https://academic.oup.com/ijlit/article-abstract/doi/10.1093/ijlit/eaag007/8673065
Every day, intimate images of real people are shared without their consent across social media feeds, phonography sites, and messaging apps. Sometimes they are stolen. Increasingly, they are fabricated entirely using artificial intelligence. Image-based sexual abuse isn't an isolated online problem. It is part of a broader pattern of gender-based violence that directly impacts the lives of women and girls. Victims can be of any gender, but women and girls are disproportionately targeted, and violence against them now makes up just under 20% of all recorded crime in England and Wales. Image-based sexual abuse and cyberflashing are now major drivers of sexual offending. They account for over half of the 11% increase in reported sexual offenses in the year to March 2025. And even that underestimates the scale because most victims never report it at all. The harm to victims is severe and often permanent. Research shows consistent patterns of fear, anxiety, depression, and suicidal ideation alongside social harms like bullying and victim blaming. These are not abstract injuries, they shape people's mental health, their sense of safety, their body image, their relationships, and their careers. And all of this is intensified by the near impossibility of permanently removing images once they're online. The law is catching up slowly. New criminal offenses have been introduced in the UK, the US, and elsewhere, including the UK's Online Safety Act and the Crime and Policing Act of 2026, which would have come into force on the 29th of June 2026. But today we'll ask whether any of it is actually working and are the regulators being ambitious enough. Today we're exploring these questions through the lens of a major new two-part article series called Fighting the Beast. It argues that criminal law, while important, is not enough on its own. The real battle has to be fought at the level of platforms and technology, and that UK regulator Ofcom needs to be far more proactive than it has been so far. Welcome to today's bonus episode of Yours Lawfully, a podcast from Q Legal, Queen Mary's award-winning commercial law clinic, in collaboration with the Technology Law Institute. In today's show, we're talking about image-based sexual abuse, what the law can do about it, what platforms should be doing, and whether our regulators are being ambitious enough. My guest is Professor Julia Hornell, Professor of Internet Law at Queen Mary University of London, where she's been based since 2000, and the Technology Law Institute Director. Julia is one of the UK's leading experts on internet regulation, online jurisdiction, and platform liability. Her work spans online safety, cybercrime, data protection, and online dispute resolution. And she has advised bodies, including the Council of Europe, and contributed to the Law Commission's work on abusive online communications. She is the author of Internet Jurisdiction Law and Practice, published by Oxford University Press, a work praised as both a rigorous academic reference and an essential guide for practitioners. She has also been a non-executive director of the Authority for Television on Demand and has spoken on this issue on BBC Radio 4 and Sky News, among others. Her latest work, the two-part article series we're discussing today, takes on one of the most urgent and underregulated harm on the internet, image-based sexual abuse. We will link to both articles in the show notes. Julia, welcome to the podcast. Thank you. So let's start by setting the scene. In your articles, you start by describing image-based sexual abuse as the beast. And I wonder at what point did you realize that existing law was simply not built to fight something like this?
SPEAKER_01Well, it is a beast, and it's also a novel beast in the sense that uh we didn't have this 20 years ago, or even really maybe 15 years ago. Um, and it's that combination of constant innovation in technology, new ways of you know, young people mainly interacting, and uh the the sort of difficulty of regulating it, which makes it a beast. And um it's the the reason why the criminal law doesn't do it is it's the problems surrounding jurisdiction, actually. So most people, when they hear what's happening online, you know, women being abused online or um intimate images being shared or deep fakes of people being created which are sexualized. Um, one of the big problems is that the perpetrators of these this type of content could be anywhere in the world. And what is the UK police doing if someone, say, just by way of example, in Malaysia is uploading this type of content or in the US? So while there's a loud cry that a criminal law should outlaw this, that, and the other, and I agree with that, but in practice the real issue is enforcement. So creating more and more criminal laws is right because we need a deterrent and we need to clearly make clear to society that that behavior is illegal, criminally illegal. But it doesn't necessarily stop it because the police will only be able to act where both the perpetrator and the victim are in the same country. And often it's also very difficult to find out who the perpetrator is. Plus, this content spreads so fast across different platforms that it's very difficult to contain.
SPEAKER_00Thank you. So um enforcement is the big issue, and uh criminal law is necessary but not sufficient. If a victim came to you today and asked whether prosecuting her abuser will get the images taken down, what would you tell her?
SPEAKER_01So the problem and up until now is exactly this that uh, you know, people repost a content and then the the victim would have to go to each and every platform to ask for a takedown. Now, platforms have an obligation to take things down. Um, that's in the Online Safety Act, in the European Union, it's the Digital Services Act, in the US it's tighter. So we have various laws which stipulate that platforms have to take down image-based sexual abuse. But the problem is that this the thing spread so fast that the victim would constantly have to issue these uh notifications to take stuff down. Then there are some platforms who will just refuse, even though they have a legal obligation to do so. Um, so it's very, very difficult. Now, the uh UK government has just passed a new law in April. It's called the Crime and Policing Act 2026. It will actually come into force on the 29th of June, which is uh at the date of recording is next week. And that provides for a very clear, uh stipulated sort of deadline of 48 hours takedown. So when a victim notifies uh, for example, image-based sexual abuse, they have 48 hours to take it down. So that helps because it gives a clear stipulation as to how quickly stuff has to be taken down. But at the moment, the victim is still left to make many takedown requests.
SPEAKER_00Thank you for really articulating the scale of the problem for victims if they want these images to be taken down and um yeah, individual having to fight a um a really big system with many platforms to reach out to. Let's turn towards um the UK's legal response to the problem in more details. You've you've mentioned the new act. In your article, you identify what sounds like a significant gap. Um, the Online Safety Act itself actually envisions proactive measures. But Ofcom, which is the UK independent regulatory and uh competition authority for the communications industry, um, has a code of practice that only mandates what big platforms are already doing. Um, so how did that gap open up?
SPEAKER_01So I guess initially Ofcom was an enormous task because the Online Safety Act, as its title suggests, has the purpose of making the internet safety. So its scope of application is extremely broad, ranging from you know, suicide, self-harm, knife crime, um, harmful um memes which children might follow through to um sexual abuse, hate speech, terrorist content, child sexual sexual abuse. So it's a long, long, broad act which covers many types of different content. So it does that. And then the other issue I think Ofcom had to face was um that it basically has to allow innovation, but at the same time restrict illegal content and content which is harmful to children. So there's clearly a tension here between, I guess, fundamental rights such as freedom of speech, freedom of expression, uh privacy on the one hand, and harm prevention or illegal content prevention. So Ofcom reacted to this tension, I think, in a very cautious way, initially focusing mainly on child sexual abuse, so sexual content concerning children, as well as terrorist content. And it was mandated by government to focus on these two areas as priority areas, and also you know, pornography websites to prevent child access to those websites. I think these were the enforcement priorities, and everything else was even if that wasn't expressly stated, but it was left more or less to self-regulation. In other words, the platforms themselves instituting the measures. Now, in relation to the abuse of, and I'm talking about adults, uh, women and obviously some men too, through intimate images, that has caused so much harm on so many dimensions that now I think the government finally has recognized that. And again, the Crime and Policing Act actually now classifies this content as priority content, so that Ofcom in the future, and we don't quite know how this will work out, will have to take a more proactive approach. But to my mind, that has been the main problem exactly, that Ofcom has been very passive, has just left this more or less to standards being developed by the platforms themselves, which were really inadequate, as we have seen with uh XAI and Croc on X, which had millions and millions of deepfakes abusing images, because obviously you can create a deep fake just by a simple image of one person, um, and then overlaying that with a sexual act, or you can uh unclothe people, etc., etc. And that has led to enormous abuse, especially of women, and especially also of women politicians, for example.
SPEAKER_00Thank you. Um, that makes a lot of sense about the very wide remit uh that Ofcom is working with. Um, you propose in your articles a new dedicated co-regulatory body separate from Ofcom to manage a unified IBSA hash database. Why can't Ofcom do that job itself?
SPEAKER_01I think Ofcom remit is probably too broad. It's kind of a super regulator. And all these issues are comparatively serious issues, right? I mean, self-harm, uh, knife crime, hate speech, etc., etc. Ofcom is also obviously the broadcast regulator. I have here in mind the model of the Internet Watch Foundation, which has always only focused on child sex abuse materials. And I think because it had that single focus, it got the buy-in from large parts, if not most parts, of the industry, and thereby was able to establish an effective co-regulatory model, as effective as can be. And I think we should have a similar focus on actually violence and abuse against women and girls, uh, in particular through technology. And therefore, I think a smaller co-regulator just focusing on this single issue would be much more effective.
SPEAKER_00That makes a lot of sense, actually. Thank you. Now that we've looked at um the regulatory bodies, um, I want us to turn towards uh the platforms that are um hosting all of this content. So in your article, you categorize platforms into five types: mainstream social media, pornography platforms, misogynist specialist sites, free speech corners, and encrypted messaging apps. Which of these is currently causing the most harm to victims and which is most neglected by regulators, do you think?
SPEAKER_01I think in many ways, I think I haven't seen this anywhere else, but I I know it's maybe a bit crude, but it's just very pragmatic to think about how we can most effectively enforce uh the law in this area. And I just thought that these platforms have obviously different commercial as well as political or ideological, if you want, objectives, and therefore they need to be approached differently. I guess maybe that's even a political point. So I think the general social media platforms have in the past at least been very willing to cooperate with regulators to make their platforms safer, maybe with the exception recently since the takeover of Twitter by X. But Facebook certainly has tried to an extent to cooperate with regulators, and so has obviously been uh Google Alphabet or YouTube. So, in many ways, these platforms, probably in terms of the, you know, simply their massive use, right? Everyone uses one of or several of these platforms, pose the greatest risk, but they also perhaps are the most likely, comparatively speaking, to cooperate with regulators. Whereas I think porn platforms come from a very different corner. They obviously have a commercial interest, they make huge amounts of money, so they're focused on money making. So are obviously the social media platforms, but the porn platforms specialize in the area of sexualized content. Plus, also they have certain commercial relationships, for example, with payment providers. So they pay again, this is probably in many ways probably the hard the most dangerous for victims because clearly many people use pornography websites, plus the reputational and trauma impact is probably greatest if these images appear on pornography websites. And then you have what I call the free speech corners of the internet and the manosphere and misogynist sites, who are obviously established often in a jurisdiction such as the US or Canada or many other jurisdictions who do not enforce the law. So it's extremely difficult to get to them in terms of enforcement. Maybe on the plus side, in terms of risk management, perhaps fewer people in terms of numeric terms will see that sort of content than on the other types of platforms. So I just thought by adopting a risk management approach, not a rights-based approach in this respect, but a risk management approach, you probably can um sort of reduce the harm with the most effect. And therefore, I think regulators should think about different tactics, different strategies in relation to different platforms. So that was the basic idea behind this argument.
SPEAKER_00Thank you. And for misogynistic special aid sites and encrypted platforms, um, in your article, you acknowledge enforcement is extremely difficult. Um, I wonder, is there a risk that tightening regulation on mainstream platforms just pushes that kind of abuse further into dark corners?
SPEAKER_01Um yes and no. I think it's very, very difficult to fight these more specialist platforms abusing women if they're not regulated, because there might be a jurisdiction which either doesn't effectively regulate or jurisdiction which in the US kind of they can rely on on free speech arguments. Um having said that, even the US now has TEDAT Take It Down Act. So there is movements in the US as well to limit this content. I think when we talk about regulation of image-based sexual abuse, it helps to think of three levels of regulation. So the first level clearly is the criminal law that would then target the users who either share or create uh deep fakes or real images of sexualized or intimate content. Then the second level is the platforms, so your social media platforms who obviously host and encourage sort of interactive user-generated content and which clearly have a mass media appeal. Um, but then also obviously other special uh different types of hosting platforms, so the platforms. But then the third one, and I think this is where the new law, the uh Crime and Policing Bill uh Act, so has had some effect in looking actually at the technology providers, so the providers of the Gen AI chatboards or other tools which allow the creation of this uh content. So it's the technology providers and the technology developers who need to also on the third level be addressed by the law. So I think it's very important that all three, sorry, all three levels are engaged, but none of them is the silver bullet. So they have to work. Thank you.
SPEAKER_00That's really clear, actually. Thank you. So we've talked about the the different platforms and um the existing regulations in the UK. Um, now the second part of your two series of articles um really focuses on existing technical measures. Um, and I wanted to spend a bit of time there because you lay out six technical tools that could genuinely change the landscape of this type of abuse. Things like hash function system, automation of content recognition using deep learning, embedding consent in pornographic images, stronger account verification, along with technologies victims can use to defend themselves and tools designed specifically to counter intimate deep fakes. We won't cover all six today, um, though listeners should absolutely go read your paper. We'll link to it again in the show notes. But I want to dig into a couple that really illustrate the gap between what's technically possible and what's actually happening. Because you speak about the gap being about trust. Um, even the best tools won't work if victims don't believe the institutions behind them will keep their data safe. Um, and right now, as you've mentioned before, asking someone to upload intimate images into a platform they don't trust is simply not realistic. So let's start with the basics. Hash matching already exists and is used for child sexual abuse material. I wonder why it hasn't been mandated for image-based sexual abuse more broadly and what's blocking implementation?
SPEAKER_01Yeah, so that's a good question because hash matching has been developed quite a long time ago in respect of child sex abuse material. But uh, and it basically works by um basically an image which has been recognized as falling into the category of say child sex abuse, um, which is then hashed, which is basically a mathematical process so that it's reduced to a number, and these are then stored in a database. The database is shared between all internet service providers. So when a user of an internet service provider or social media tries to upload any image, every image, um, it will be hashed. And if the hash function matches, then there is a very, very, very high degree of likelihood that it's exactly the same picture, or maybe a similar picture but with slight modifications, but essentially depicting exactly the same. So the internet service provider can simply block. That image from being uploaded. So that's how the hash function technology works. But the key here is that we need more than just this software or this mechanism. We actually need a sort of environment or we need a database, then obviously all the internet service providers need to subscribe to this database and implement a hash matching system. So hash matching has been discussed quite early on, I can't remember now, maybe 10 years ago or so, by Meta, Facebook, who suggested that women can upload their intimate images so that it can almost kind of create a preventative database. But of course, that wasn't very successful. Why? Because obviously, who would upload their intimate images into some sort of database hosted by Meta? And then similarly, Pornhub was also obviously pressurized to do something similar so that there are no non-consensual images on Pornhub, because obviously Pornhub has a huge amount of non-consensual images. It has tried to clean up its act, and as part of that, it has sort of set up a sort of hash database which it has outsourced uh to another provider. But still, I think women wouldn't necessarily trust this uh method. And also, again, it's completely compartmentalized. So women or victims, you know, men too. So I shouldn't just talk about men, but whoever is a victim would then have to upload their photograph to each and every porn platform or the outsourced entity. Again, that's not going to work. So in the UK, the Safer Internet Center has set up a system actually where people can upload images. So that's a trusted body. But they're not very well funded. They need much more money, they need much more resources, and they also need the institutional power to persuade all the internet service providers to actually join that system. Obviously, some have joined, but not all of them. So I think that was the great success of the Internet Watch Foundation that in relation to child sex abuse material, they have managed to get the industry around the table by and large and um cooperate in the system. And I think we need to build something similar for image-based sexual abuse. So we need internet service providers to subscribe to it, we need them to pay for it, I guess ultimately, and we also need them to share that database. So I guess the way with uh this works is that they um the the co-regulator would license the database against a fee, which then allows that company to use that database. The problem here, I think, is with the big tech companies, they want to do everything themselves. So the difficulty is to network them in such a way that they buy into the system. Now, the good news is that the uh Crime and Policing Act from last April basically envisages such a system by giving the I think Secretary of State the power to kind of set up such a system. That obviously hasn't been done yet, the act isn't in force, but it's envisaged in that. So I think it's very much now a question of setting up a suitable system, hoping that the big social media companies and also the porn companies sign up to this.
SPEAKER_00Thank you. I mean, it's making me appreciate um how there's a real need to design both the tech measures and the law for you to be being in contact with the victims so that what is actually happening to them is understood and um tech measures and the law can be designed to meet their needs. Um, otherwise it can fall short quite in a quite big way. Um and and linked to that, uh you also talk about uh, of course, consent um is um is key and difficult thing for legislators to regulate. And in your article, you talk about embedding consent directly into content as a technical standard, which is really a fascinating idea. I hadn't come across it before. Um, can you walk us through what that would actually look like and what stops platform from simply not implementing it?
SPEAKER_01So my thinking here was this again, very pragmatic. I was thinking, well, in relation to um you know non-consensual um intimate images or image-based abuse, the two elements which stand out is first of all the lack of consent, and secondly, obviously the intimate nature. Now, large language models and um AI tools, I think, are now pretty good at recognizing this type of content. Um, they're not foolproof, but they're they're pretty good. So that leaves us to the question of consent. And what I was thinking here again, that I assume that the greatest in terms of risk management, the greatest harm stems from social media companies and the big porn platforms, because everyone uses them. So I thought if we could find a technical way of embedding consent and use AI to recognize what image is intimate, then a lot of images could simply be blocked from uploading. So even if the image hasn't been hashed yet or hasn't been recognized as image-based sexual abuse, we could use this proactively to ever stop such an image appearing on the social media side. Because clearly the hashing technology requires the victim to find the image and notify the company that this is an image-based sexual abuse, and then it would be put into the hash system. My system of this idea of embedding consent would be to recognize that something is an intimate image and it's non-consensual before it's uploaded the very, very first time to stop it right in the tracks. So that was my original idea. Now, I acknowledge that consent and consent to um pornography and pornography production is a very sensitive and disputed issue. And I totally agree with this. So, but the article looks at a much narrower aspect. Literally, the article assumes, and this might well be, you know, uh it's just one part of the reality. Um, so basically, what I want to prevent is where the the person depicted has consented to the making, the creation of this intimate image, that those pictures, and I assume that those pictures uh then get uploaded without consent. So consent happens actually at two stages, right? Consent happens at a creation, and consent happens at the sharing stage. Now, obviously, if porn producers would probably, at least formally legally, consent to both the production and the uploading and the sharing, I guess that's what you do as a porn producer. And as I say, that raises other issues about the validity of consent. In extreme cases, women are trafficked, they're forced to do this sort of work. Is that real consent? That problem, my article does not deal with it, it just doesn't deal with that. That's another level, another issue. My issue is just where images, where there's no consent whatsoever, and the image is uploaded without a knowledge of the person uh in the image. So, how would you embed consent? Well, if porn uh performers could, and there might be better ways, I'm not a technical person, but my initial thinking was this if a porn producer would produce a short consent statement before the video starts, and if he can also verify that this is a real video and not a deep fake in itself, then surely that would do the job, right? And you embed that consent then in the actual picture. If that embedding is missing, then again the image will not be uploaded onto the porn platform. So that in a nutshell was my thinking.
SPEAKER_00Yes, and I suppose the challenge is getting all the platforms on board with um implementing this as a technology, as well as the technology keeping up with the technological changes, as you said, so that um deepfakes can be uh found out and um and and signaled. Um, but fascinating idea, really important discussion. I'm conscious of time. We could talk about this for for much longer, but I really want to ask you about a bit more about jurisdiction, which I know is um one of your many expertise. You describe in your articles offcomes business disruption measures, such as blocking platforms, cutting off payment services as a kind of secret weapon against foreign platforms. Um, so I want to ask you how much of the beast can actually be tamed that way?
SPEAKER_01So I think the enforcement challenge is probably one of the biggest challenges in this area. So, for example, Ofcom has been very proactive in its enforcement. I mean, obviously, it has limited resources. It started with platforms which provide pornographic content to minors without age verification. And some of them simply complied and did better age assurance or age verification. Others simply ignored that. So, for example, in the news it was reported that 4chan, which obviously allows users to share pornographic uh images of all sorts, you know, um, that they just said, well, we're covered by free speech in the US and we're not interested basically in engaging with the regulator. So the jurisdictional issue is that obviously the internet can be accessed from anywhere. Many of these platforms will simply be uh located outside the jurisdiction, which makes enforcement really difficult unless there is sort of enforcement cooperation between criminal authorities or regulatory authorities across the globe. Now, obviously, in the US, we've seen, especially with the current Trump administration, has been a huge pushback at content regulation on social media. So there is no cooperation whatsoever. On the contrary, they're actively undermining any type of enforcement. So given the geopolitical situation, also the jurisdictional challenge, what can Ofcom do? Well, it can try and use local intermediaries within the jurisdiction in the UK to enforce uh the regulation. So, for example, payment providers or app providers, so blocking you know certain apps, or uh website blocking, of course. Now, website blocking is not particularly effective, it can easily be circumvented, uh, VPNs, etc. Payment has some effect, so there was this scandal about um Pornhub hosting, um, child sex abuse material, Pornhub posting, non-consensual intimate abuse. Um, and the payment credit card companies got hold of this and basically then said, Well, we stop providing you with payment services, which obviously caused a major hoo-ha for performers on Pornhub. So they were unhappy because they couldn't get any payments. So obviously, Pornhub had to clean up its act. Whether it's done so effectively, who knows? Now, Ofcom has not yet, as far as I'm aware, used uh these uh orders. So obviously Ofcom would have to go court and then get a blocking order or get an order against a payment provider or advertisers, so those intermediaries uh who are in the UK. Um, but it's definitely in the Online Safety Act, so that power is there, and I think it will be used as well. So, yes, it is probably another another sort of um arrow in the bow of um enforcement.
SPEAKER_00You've talked about um the TIDA in the US, um, which effectively pierces section uh 230 immunity for image-based sexual abuse. And I wonder um how significant is that globally? Does it change the calculus for platforms operating internationally?
SPEAKER_01So various US states had criminal laws on uh various forms of image-based sexual abuse for quite a while. Most of them now, most states have that, but there was no federal criminal offense until TEDA was enacted. So that meant basically users all over the world, including the UK, when they complained to US headquartered platforms, the platform could just say, well, we're covered by Section 230 of the Communications Decency Act, which means we're immune from any liability. Now, Section 230 does not give immunity from federal criminal law. So the enactment of TEDA means that there is now federal criminal law, and that means if a platform is notified, they cannot hide between section 230. So that has actually a positive effect, not just obviously on victims in the US, but also on victims in the UK. So in the past, victims, if they had copyright in the images, they had to rely on copyright law in order to get the images down. But TIDA effectively has uh removed that.
SPEAKER_00So that's a that's a pretty big change. Um great, thank you, Julia. I have one last question. Um, if you could change one thing, one law, one platform policy, one technical standard, that would make the biggest difference to victims right now, what would it be?
SPEAKER_01I think there are two. So I'm cheating here. I think the first one is what I said earlier. I think if we could have a co-regulator like the Internet Watch Foundation, which is just focused on abuse, sexual abuse, mainly of women and girls, but basically maybe sexual abuse, enabled by technology. Because I think technology has hugely changed the nature of pornography, but also sexual abuse. And I think if there could be a regulator of the respecting freedom of expression, privacy, but we would just focus specifically on image-based sexual abuse and is properly funded, that would be a real game changer. And if that regulator could then also assist women and provide victim support, that would be great. And the second one goes back to something I've said earlier. So at the moment we have criminal laws and they've been again expanded in the uh crime and policing bill, which is great. Um, we have platform regulation, the online safety act, which Ofcom looks after, that's great too. But I think the third level, the technology regulation itself, so in other words, forcing Gen AI providers to better build guardrails against this type of content through various technological means and standards, that is the next step. So the Crime and Policing Act contains a specific provision which says that a Gen AI provider or developer or provider of, say, a chatboard, which has the purpose of image-based sexual abuse, that would be a to provide such an app would be a criminal offense. But it's not entirely clear what that means with the purpose of. I mean, most of these Gen A Gen AI tools, and that's the beauty of it, they can be used for all sorts of people, you know, purposes. They can be used to write a CV, they can be used to create image-based sexual abuse. So I think we need to develop standards and impose those standards on the tech providers. So they have to make sure that their technology can't be abused in this way. And so we have to develop the technical standards to make sure this doesn't happen. And these guardrails then have to be implemented through regulation.
SPEAKER_00Thank you. It's hard to pick one when the need is so great. Um, thank you so much, Julia, for all your many insights today. So, what have we learned today? Um, criminal law matters, but it will never be enough on its own. The volume of abuse is too high, and the perpetrators too anonymous, and the borders too porous. The real fight happens at the platform levels. Um, and Ofcom right now is probably not being ambitious enough. We've looked at why. Um, the Online Safety Act gives it the power to demand proactive measures, and um, hopefully in the future, it will be using it more. The technology to do better already exists: harsh functions, automated content recognition, deep fake detection. What's missing is the regulatory will to mandate it, and perhaps, as you've mentioned, a co-regulator. And that behind every statistic is a person whose images are circulated online right now, today, with very little recourse. So Julia Hornell's call to action is simple: being strategic, being proactive, and let's hope the regulators are listening. And with that, all that's left to say is a massive thank you to Julia Hornell for joining me on today's episode. Thoughts, feedback, and questions are all very welcome. Thanks for listening.