Good afternoon, and welcome to the Women and Equalities Committee. Today we are holding an evidence session on misogyny, the manosphere and online content. I am really grateful that we have Laura Bates here, the author, writer and activist. Thank you so much for offering your time and expertise to us today. I should also say that we are joined by our colleague from the Science, Innovation and Technology Committee, Emily Darlington. Thank you very much. We will be asking questions, but if there is anything that you think, “Actually, I want to go into greater depth but we don’t have time today,” or when you leave you think, “Ah, I wish I had said this,” please send it into us in writing, because we want to make full use of your expertise and knowledge today. You have recently written a book on AI and misogyny. What are the dangers that AI poses to women?

There are many. That does not mean to say that we need to be anti-tech. It means that we need to recognise that perhaps the biggest danger of all is that at present the perception of risk from AI tends to focus on potential future dystopian threats. Will it take everybody’s jobs? Will it wipe out humanity and take over the world? While those are understandable and important issues, there is not much recognition of the way in which it is already affecting women and girls’ lives today in the here and now. That is both in the form of targeted tools deliberately created for abusive purposes—for example, deepfake apps and technology—and in shifting societal norms and perceptions around sexual violence through sex tech such as sex robots and AI companions. There is also the inadvertent replication of existing bias, which is rife in AI tools that are being used already across global financial services, for example, where AI is being used to determine credit scores. Already, we have a $17 billion global gender credit gap, so this is a significant existing problem that risks being exacerbated. There are similar issues around healthcare, for example, with tools being created that are more effective at detecting things in male blood samples than in women’s. We know that 40% of UK firms are already using AI tools in their recruitment processes, but again, AI recruitment tools have been proven to explicitly discriminate against the CVs of women and marginalised groups. That is not to say that they have been built to do that deliberately, but rather that they use existing huge datasets that they are trained on, and they then regurgitate that in guessing what the likely best answer is. They replicate—if you like, they consume and vomit back at us—our existing racism and misogyny, and risk exacerbating them through a process of concentration, because they are by definition designed to eliminate outliers. There are a number of different ways, I would say.

That is a great—if depressing—way to start.

Thank you for joining us today and thank you to the Committee for letting me join in. I wanted to start by saying that your book “Men Who Hate Women” was for me quite revolutionary in allowing me to see the difference in what I thought were isolated incidents, but also to see the radicalisation of young men and some young women online. I wanted to get your assessment on the effectiveness of the Online Safety Act that has just come into force here in the UK, and whether you think it is able to deal with online misogyny, particularly illegal and harmful online misogyny.

It is difficult to say yet because the proof will be in the pudding, if you like. What is going to be very important is how it is implemented. Every pre-existing example that we have had of efforts to regulate has been a failure, particularly around social media. What we have seen is that where there is a heavy reliance on assumptions that social media companies will comply, that tends to end in failure. I have concerns that that may be an issue here. It is a huge amount, in terms of the sheer scale of material, for Ofcom to tackle and regulate. For example, if we look at deepfake issues alone, we are on track to see 8 million new deepfake images created by the end of this year, and a 2,000% increase in referral links to websites hosting non-consensual deepfake pornography across social media. There is a question of scale and how that will realistically be implemented. There are also concerns about a reliance on the idea of waiting for harms to occur and then retroactively trying to take content down, which again will rely on co-operation by big tech. I would like to see a more proactive stance in terms of thinking about prevention and particularly thinking about tweaking algorithms. That is within the scope of the Online Safety Act, but it is then on the regulator to actually really be prepared to take action in enforcing that. It is not something that we have seen successfully done anywhere in the world to date. I would also say that valid concerns have been raised about certain loopholes, particularly around a lack of focus on smaller websites. Understandably, they are often perceived to be perhaps less of a risk, but the reality is that when we are talking about this particular form of harm, we often see concentrated local issues. For example, deepfake pornographic images of local women are made to order, with women named and their jobs, education or addresses listed. Often, in terms of victims coming forward, the most intense harm is actually localised. The other issue is around thinking about intersectional issues. There are not necessarily provisions tackling these issues in so far as they pertain to women of colour. This is particularly the case for women who, for example, might be harmed by a deepfake image being created that shows them without a head covering if they normally wear one, or with a man that they are not related to in situations where that would be considered problematic. There are valid concerns, but it is also about seeing what happens and how well it is done.

You have already answered my second question about harmful but small platforms, and the role that they play and whether there is enough oversight of them. My last question is one that can be about hope, or maybe despair. Obviously, all these platforms rely on algorithms, and those algorithms are driven by AI. Do you see any work through your research where these algorithms are actually being programmed to weed out this material—to identify AI-generated material and either list it as such or to weed it out of what is being pumped into our timelines or anybody else’s, including children?

Yes. We know through hash sharing, for example, that it is possible for that technology to be used in such a way. For example, victims who might have been coerced into sharing an intimate image can choose to create a hash of that image to prevent it from being uploaded. It is then reliant, of course, on platforms, social media websites and porn websites choosing to sign up to and comply with that. There are examples of really brilliant feminist work being done within AI. For example, feminist researchers have harnessed AI in South America to look at unjust patterns of sentencing decisions in court cases. There are examples of AI tools being developed that can be used to police and identify bias in other AI tools. We have not really seen any effective example of the algorithms that are used to drive the most extreme forms of harm on social media being regulated. This is something that Frances Haugen—the Facebook whistleblower—focused on in her testimony to the US Senate. She talked about the connection between algorithms designed to drive engagement ruthlessly above all else, and the spread of misinformation, really harmful content and severe real-world harms. This is also something we know from YouTube whistleblowers. We know that these algorithms are not designed to promote the most effective, highest quality, or most relevant content but that which keeps people watching for the longest, which studies show is increasingly extreme content. You are right to identify that the algorithms are themselves one of the biggest sources of the problem. They are also one of the places that have been hardest to get to in terms of regulation. Even just greater transparency around that would be a really important first step, in my view.

Fantastic. If you were to make a suggestion to Government about the kind of regulation that could deal with it, what would be your advice on that?

We would need significant enough transparency to be able to identify the links between algorithms and harmful content being accessed. What we are seeing at the moment is that platforms simply do not comply, even with their own moderation policies. Where you have a platform like TikTok, for example, which claims that it will not allow misogynistic content but has displayed Andrew Tate’s content until it has been viewed over 11.4 billion times—more than the number of people on the planet—that is a policy that is not fit for purpose. What we are seeing at the moment is that there is simply no kind of adherence. I would suggest that in the light of that context, there needs to be really significant regulation and proof of compliance with that regulation.

I would certainly concur with that given that when X was in front of the Science, Innovation and Technology Committee, I read out my death threat and that account is still up, as is the tweet that was sent.

I am so sorry. Thank you, Emily.

It is a really significant issue. You could be forgiven for not realising that if you have looked at the political and media focus on deepfake harms, because it too often focuses on the idea of potential political disinformation and its threat to democracy. Of course, that is a hugely important issue, but we know that 96% of all deepfake images are pornographic in nature, and that 99% of those feature women. Significantly, where deepfakes exist, they are already now devastating the lives of women and girls. It is a society-wide issue. It is affecting local women who are targeted on specific made-to-order websites. It is also having a significant impact on women whose lives are by necessity in the public eye to a degree, such as journalists and politicians. For example, 40 MPs were targeted in hundreds of deepfake images that were hosted on a website that had 12 million views according to a recent Channel 4 investigation. We know from an investigation in the US that female members of Congress are 70 times more likely than their male counterparts to have experienced this form of abuse. It has a devastating impact—I am sure I do not need to tell you here—on the lives of people who are trying to do good things and trying to play an important public role in society. It has a significant stifling impact, I would suggest, on young women and girls thinking about whether they would feel able to take on a role like that in the future. Across society, there is a significant impact. If we want to talk about the threat to democracy, I wish we would talk a bit more about what that looks like now in terms of female politicians, instead of potential future risks. When it comes to schools, I would say this is an invisible epidemic. The problem with schools is that very often—through absolutely no fault of the hugely committed and passionate teachers and staff who are working so hard to tackle these issues—they are about five years behind in terms of recognition of what the issues are and how to tackle them. Five years ago, I wrote a book called “Men Who Hate Women” about the radicalisation of young men and boys by extremist influencers online. Schools are now starting to talk about those issues, which is welcome and a really important step forward. What I am seeing now among the young men and boys I work with in schools regularly is that deepfake pornographic image creation is a really significant problem. But when I talk to schools and do teacher training around these issues, there is a sense of shocked disbelief and a sense that, in many cases, teachers are not even aware of this technology. It is having a very significant impact on young women and girls, who are often aware that these images of them exist. In 2024, the first story hit the headlines of a school in the UK where this had happened. Around a dozen girls were affected and very little action was taken. It was a nearby boys’ school in that particular case. What is really significant is that in every UK case I have investigated where this has happened, schools have paid thousands of pounds hiring outside PR firms to focus on damage and reputation management. In terms of focus on supporting those girls, working with frontline services, and taking action, nothing seems to have happened in the cases that I have seen. In fact, I wanted to read to you a quote from a parent of one of the victims in one of those cases. She told The Times, “This has been really hard for our daughter...To find out that these videos had been created of her and… circulated was a horrible shock. For her to see, seven weeks later, that no one has been disciplined and that she has had no form of apology is even harder…she is…coming to the realisation that this is how it is going to be—something that she will just have to put up with. Not something I ever imagined my daughter, in 2024, would have to accept.” It is happening; it is significant. Female teachers are affected, which often goes unnoticed. We are seeing female teachers having deepfake images of them created and circulated, and schools are just not equipped to tackle it even if they know that it is going on. My suggestion would be that this is the next big sexual violence epidemic already occurring in schools, and the biggest problem at the moment is that people do not even know it is happening.

In the cases that I am aware of, investigations by police are ongoing. I am not necessarily suggesting that criminalisation of under-age boys is even the solution in this case. Ideally, with all these issues, we want to think about prevention, education and regulation. It should not be the case that a 12 or 13-year-old boy can easily and immediately access a free tool on his phone, or by googling, that will give him the opportunity to create these abusive forms of content in the first place. In terms of school responses, I am not aware of any instances where schools have taken effective action. We tend to see a repeat of the failings we have seen in some schools around intimate image abuse previously, where we saw girls being coerced into sending images of themselves that were then spread around. What we tend to see is girls experiencing a huge societal backlash, with girls themselves sometimes being punished by schools for having taken the photos. There is not much disciplinary action being taken towards boys and not much happening in terms of prevention. Invariably, girls ending up dropping out of school or education altogether as a result.

To my knowledge, there is no statutory guidance for schools about what to do on this specific issue. I may be wrong about that, but I have had a look. What we need is very clear guidance on exactly how schools should respond in these cases, particularly around a duty of care for survivors, should there be an ongoing investigation by police. Even if disciplinary action is not immediately taken, it does not mean that support for survivors cannot be implemented. It does not mean that work with local frontline organisations, for example, cannot be started. This also needs to sit within a broader framework, where schools need better statutory support on a range of issues around how to deal with sexual violence more widely. This is also the case more generally when we think about regulation and tackling all these issues. How are we expecting Ofcom to be effective in taking action on a form of abuse that is affecting tens of thousands of women and girls in the broader context of a justice system where less than 2% of people who come forward to report a rape to police see it result in a charge or summons for a perpetrator? It is also important to recognise that difficult context of a wider system that is not fit for purpose in these cases.

There is a lot that parents can do. First, I always say to parents, “I recommend giving yourself the experience of what the online world is like for your child.” If you have a 13, 14 or 15-year-old boy, set up a TikTok account, tell it that you are a 15-year-old boy, and see what kind of content it promotes into your feed. We are living in a unique moment in history that we do not really discuss very often, but it is very significant—it has never happened before and will never happen again—where a generation of non-digital natives are parenting and educating a generation of digital natives. That creates really significant challenges. An awareness of what the online world looks like is a really good starting point. If you have seen the kind of TikTok videos that are promoted to young boys’ feeds, you will feel a bit more confident to talk about them. You will understand some of the language that your children might be coming home and talking about. Take a look at some men’s rights activist pages on Reddit, take a look at some big Instagram meme accounts, and have a look at those TikTok feeds, because the reality is that there is nothing inherently misogynistic about this generation of teenage boys. They are growing up in a world where when you start a TikTok account and tell it you are a teenage boy, on average it takes less than half an hour before the first piece of extreme misogynistic content is promoted to your feed. You do not have to go looking for it, and most boys are not. This is an algorithmically facilitated form of mass radicalisation. Go on to YouTube and put in a video about women or something quite benign and see what kind of videos the algorithm suggests that you watch next. Some 70% of all content consumed on YouTube is made up not of the videos people go looking for, but by the videos served up to them by that algorithm, which as we have already said, prioritises extreme content and increasingly so. There is a lot that parents can do in terms of letting themselves become aware of the problem. Secondly, I would suggest communication. It should be little and often, not one big terrifying conversation when they are already 16 or 17, but starting from an incredibly young age. Give them permission to recognise gender inequalities and stereotypes in the world around them and to challenge them, whether that be in children’s toys, magazines, clothing and books, or in advertising and movies. Pointing it out and talking about it demonstrates to them that it is not just normal, and that it is okay to disrupt, challenge and question it. By doing that very gradually over time, you are giving them the tools later on to have a form of source scepticism when they come across claims online, on social media. Internet literacy is something we can all talk to young people about from a really young age. These are all things that parents can start doing little and often. The last thing I would say is that far too much pressure is being placed on parents to fix a problem that they have not created. There is an obsessive focus on what parents should do and whether they should ban their child from smartphones or from social media, which puts them in an impossible position. Do I choose to ban my child from this access to tools that will be vital for their future and for future careers but protect them from harm, or do I allow them to access a world that is important for them socially and developmentally but know that they are in a place where they could experience harm as a result? The answer for me, very clearly, is to prevent the harms so that parents can enable their children—as is their right—to be digital citizens and to access those tools without facing extreme harm as a result. For me, it is part of a focus in which we would do almost anything to avoid looking at regulation and tech companies as the solution, but which just shunts an impossible burden on to parents instead. This is not something that they can fix really.

It is a step in the right direction, of course, and I welcome that step, but it is deeply disappointing that the Government have rejected many of the proposals made by this Committee, particularly around the possession of such images, which would have brought them in line with child sexual abuse images. That would have sent a very clear message. I cannot see any justification for the legalisation of owning these images. Other proposals—again, you will know these better than I do—were about extending prosecution time limits to account for delayed discovery, particularly around creating fast-track civil processes because we know that the justice system is not fit for purpose on these issues. That would have made a real tangible difference to survivors. We also need to think about longer-term funding for frontline services that are doing an incredible job of supporting survivors who experience these forms of harm, but are very much under-resourced and underfunded.

I feel a bit starstruck, to be honest, because I have been really inspired by you ever since you started the Everyday Sexism Project, so thank you for all your work. It is brilliant. I am going to ask about an area I know nothing about, which is virtual reality and gaming. I have two sons, but they are older teenagers now at university. When they used to play online, it was Minecraft, and it all seemed really innocent and lovely, and I really enjoyed watching them build their worlds. Now we have Roblox and all sorts of other things, and the metaverse. Can you explain to us what the dangers are of those virtual reality spaces for women and girls, and are there any safeguards in place?

As you have identified, there are a number of different large-scale virtual worlds that people can interact in. When we talk about metaverses, people tend to assume that we are talking about the metaverse envisaged by Meta, which it has spent tens of billions of dollars creating. That is really a 3D virtual reality future world in which it not just hopes that we will be able to meet and interact in an embodied version of social media, but very much has ambitions to see us sitting in virtual classrooms, meeting in virtual boardrooms, going to virtual events, shopping in virtual shops, and so on. The reason it is really important to contextualise it in that way is that we are not just talking about an online gaming space that does not really matter because people have the opportunity to choose not to enter it if they do not want to. We are talking about the beginnings and foundations of what might very well be a future world in which many of us have very little choice about living. That is why it matters that we get this right now. If we look at early iterations like Roblox, what we have seen is repeated instances of extreme harm, abuse—particularly of children—grooming, and cases that have escalated into offline real-world harms. In the metaverse, I would suggest that the picture is similar. When I researched my new book “The New Age of Sexism” I spent days in the metaverse. I was there for two hours after entering using one of the company’s 3D headsets before I first witnessed somebody being sexually assaulted virtually in front of my own eyes without having gone looking for it at all. Sexual harassment and abuse were just a constant feature. When I asked other women around me after witnessing this, “Have you ever been assaulted in the metaverse?” The response was universal, “Yes, of course, almost every day. Everybody has, it’s just the way things are.” This is not just anecdotal evidence; the Center for Countering Digital Hate recently did a study on this, and it found that every seven minutes, on average, users are exposed to abusive behaviour in the metaverse. To come to your question about regulation and what is being done about it, Meta would of course point to its moderation tools. But the CCDH study uncovered 100 violations and tried to report them all, but only 51 of them could even be reported through Meta’s moderation system because it has predetermined categories. If something does not fit into one of those, you cannot report it at all. But it reported 51 of them, and not only was no action taken in response to any of those reports but actually there was not even any response to any of them. It is not an exaggeration at present to say that the metaverse, which again lots of people think of as a future potential space, already exists now. In my experience, it is already the site of repeated harms. This is something that has plagued it since its inception. Researchers have reported virtual rapes and virtual sexual assaults. UK police at present are investigating the virtual gang rape of a girl under the age of 16 in the metaverse. A senior officer familiar with the case explicitly made a point of saying that the girl had suffered psychological trauma similar to somebody who had experienced offline sexual violence. That psychological impact is key because there is often a dismissive attitude towards these forms of harassment: “Just take your headset off,” or “Ooh, I got shot in Call of Duty, but it didn’t really hurt.” But this is of course a completely different thing. You are not entering the metaverse fully aware of the fact that you are going to be playing a game in which you are going to shoot people. You are not expecting to be sexually assaulted or harassed. Really it does not matter how similar these harms are to separate offline forms of sexual violence; the fact is that they should not exist and should not be happening. It becomes particularly significant when you look at the very quick escalation of haptic wearable technology. For example, you can already order full-body suits that will have hundreds of receptors, which will enable you to pick something up in the metaverse and have it feel real. When a gust of wind blows towards you in the metaverse, you will feel it. In a similar way to some 4D cinema, this technology is escalating at pace. It will be a very short time before the experience of being virtually sexually assaulted in the metaverse will feel very little different from an experience of physical sexual assault in the real world. Again, that is just another reason why it is very important that we tackle this now before things escalate further.

I was reading that the CEO of Roblox said, “If you’re not comfortable, don’t let your kids be on the platform.” That does not really seem good enough to me. Would you suggest that there needs to be age verification and some kind of lower age limit put on, or any other things put in place to stop anyone under 18 going on them?

Yes, there should be. At present, five-year-olds can have a Roblox account. Research repeatedly suggests that even with the highest parental settings, it is still relatively easy for abuse to happen on the platform. I believe around 60% of US children have a Roblox account. The scale of what we are talking about here is too significant to be reliant on the assumption that parents are aware enough of the inherent dangers to take that step. It is entirely understandable for parents to think, “Oh, this looks like nice game. I’ll let my child play on it.” It should be incumbent on the tech companies themselves to make it safe for their users, not just shunt that responsibly on to parents. Yes, parental controls are important, but there is very little prioritisation of that. After Facebook launched its first 3D headsets—its first metaverse tools—it was three years before it created its first set of parental controls to be used for them. This is really significant because the NSPCC estimates that while 150 apps, games and websites were used to groom children online between 2017 and 2023, where the means of communication was known, 47% of online grooming cases took place on Meta-owned products. This should be treated as a public health concern. It should be something where we consider that the safety processes and protocols should be in place before people enter these spaces, in just the same way that we would all accept that—without any consternation—for a real-world space, like repeatedly delaying the opening of the Co-op Live arena in Manchester because it had not yet met the necessary safety protocols for audiences to come inside. There is no reason why we should be treating this any differently.

I have a question on the difference between how offline and online spaces are treated. The NSPCC found that 41,000 online grooming crimes had taken place against children between 2017 and 2024. That was an increase of 89%, and it is only increasing. Some 81% of the victims were girls. The NSPCC stated, “Perpetrators typically used mainstream and open web platforms as the first point of contact with children,” as you were saying, “This can include social media chat apps, video games and messaging apps on consoles, dating sites, and chat rooms. Perpetrators then encourage children to continue communication on private and encrypted messaging platforms where abuse can proceed undetected.” If we are talking about accountability here in an offline setting, that setting would be accountable for the safety of the people using it. Why is the online space unaccountable for the safety of its users? Can we be doing more to trace the perpetrators of these crimes in the online world, whether that is through their payment systems—with Visa, Mastercard, and Apple Pay being more willing to share data and information—or is it that we need better identification and verification of anonymous accounts?

It is an excellent question, and it is a complicated one. The simple answer to why we are seeing this happen with impunity when it is a public health catastrophe that simply would not be tolerated in an offline venue is because of the incredibly powerful tech lobby. The tech lobby has been extremely successful in convincing politicians globally that this is something that they should not be held accountable for, that they are platforms not publishers, and that this is something that is too big, too difficult, and too unwieldy for them to be expected meaningfully to attempt to moderate. We simply would not accept that argument for any other sector. We would not accept a multinational food company saying, “Well, millions of people consume our foods around the world every day, so you can’t expect us to prevent a few of them dying of salmonella. It’s just one of those things. It’s a matter of scale.” You simply would not have a company saying that. At the moment, the argument that is being made for accelerationism—the very trendy concept in Silicon Valley right now that there has to be no regulation in order for unfettered development to take place—is that any Government brave enough to regulate will lose out dramatically on innovation because those companies will go and innovate elsewhere. The problem is that you could have made that argument throughout history about any form of innovation. You could have said, “Well, okay, we’ll say that car companies can make cars in our country and there don’t have to be any regulations or safety standards met.” You might get a lot of people coming to make cars in your country, but I am not sure they are the kind of cars you would want people in your country driving. There are huge trade-offs being made in terms of ethics and safety that will have a significant impact on us all later down the line. There is a moral and ethical obligation to think about this in terms of the wellbeing of citizens, and in terms of the tools that are going to be effective, robust and profitable. Actually, ethical and safety standards are important for those things as well. We cannot really talk about this without recognising the very significant and worrying influence of the biggest tech firms at the moment on the Government of the United States and their presence in the Oval Office alongside Donald Trump. For researchers in my position in the UK, it is particularly worrying that they have witnessed suggestions that that has had a significant influence on the decisions made by our own Government on these very important issues. One example is the AI Action Summit in Paris earlier this year. Very notably, the UK and US Governments refused to sign a pledge suggesting that AI should be ethical and safe, even though 60 other countries signed it. Another example is reports around the time of negotiations for a trade deal with the US, when it was suggested that Government sources might be prepared to consider watering down certain provisions of the Online Safety Act to secure a good trade deal. There have been some worrying things about the impact that that is having on attempts to regulate.

Why would any Government not want AI to be ethical and safe?

What they would argue is that any kind of restriction and regulation is anathema to the tech companies creating these tools. They will therefore go to countries where they do not have to jump through such high hoops at development stage to create their tools. There is a sense of panic and a fear of being left behind. This is being discussed as the new space race or the new gold rush, but it is absolute madness and very self-defeating to think that the most quickly-designed tools—absent safety regulations and ethics—are going to be the ones that are the most effective, the most productive and the most profitable in the future. It is also just a complete flagrant lack of responsibility towards citizens to suggest that what is essentially profit should be prioritised over people. So many of these tools are not being designed with ethical aims in mind; it is very much about profit. It is about increasing the existing obscene wealth of a very small number of men at the expense of the wellbeing of human beings, particularly women and marginalised groups. It is worth saying that we have seen this happen before. We do not need to guess what is going to happen. Exactly what has happened with social media is the perfect microcosm of what we are now about to see happening on an exponentially greater scale with AI. We saw that voices of concern were raised right at the very beginning. Often, they are the voices of women of colour, who are the canaries in the coalmine who say, “This isn’t ethical. This isn’t safe. There is abuse. There is bias.” Their voices are ignored and the platforms then burgeon to a degree where they are now inextricably intertwined with many of our day-to-day lives, our careers, the way we get our news, our social lives, and so on. At this point, it is difficult to choose to take yourself off those social media platforms without suffering significant issues as an individual, so we are reliant on trying to get companies and platforms to regulate. But now that they are so enormous, and frankly so profitable, there seems to be very little hope of really getting that retroactive regulation in any effective foundational system change way. Now we are at the beginning of that same pattern happening all over again with AI. We have this moment of opportunity to say, “This time around, shall we act now? Shall we demand that there is common-sense regulation and safety protocols that have to be cleared before products can be rolled out to have an impact on the public?” If we do not do it now in that order, we will see a repeat of exactly what has happened with social media, which is companies essentially using trauma, abuse and discrimination against women and marginalised groups as a convenient building material. Q130       Samantha Niblett: I am the founder of Labour: Women in Tech. I have a background in data and technology. I have worked for companies before they were regulated and then afterwards. Before they said they did the right thing; afterwards it looked slightly different because they had to do the right thing. You have possibly even answered my question in everything you have said already. I am personally being lobbied by Meta to support its methods for age verification, which to me comes with alarm bells having read Sarah Wynn-Williams’s book “Careless People”, because we know that it wants to regulate for its benefit. There are lots of recommendations such as, “Don’t just let big tech have your ear. Don’t believe everything they say. Government need to be regulating.” Perhaps we can have a separate conversation about this, but it would be really good to understand what it should look like and what would be perfect. Something you have said today, which has been quite astonishing—I have not thought about it this way—is that a new world is being built, and it is online. We are not building it; they are building it. Perhaps it is not for now, and perhaps you can share information later, but it would be good to understand what we should actually be doing. How can we regulate it and ignore the beats if it is just too hard?

Yes. There are some really good proposals around what that regulation might look like. The European Commission in particular has created a framework that you could think of as comparable with the Istanbul convention. That is so powerful because there is this sense that if it is transnational, how do we legislate in a way that is transferable to different global contexts? Well, that is how you do it. That is a really good model to look at, in my view. There has also been a report released just this week by tech leaders in California, commissioned by the governor of California, to look at what realistic regulation might look like. Again, they have suggested common-sense proposals for transparency and independent scrutiny of AI models before they are released to the public or have an impact on the public. It is not rocket science. It is doable. We do it with other complex international issues. It is just the case that tech companies would very much like us to think that it is impossible. If we think of the ways in which failings have happened on text-based social media websites and other websites—where the evidence is there to search and use tools to find and retroactively to examine—imagine the ways in which that is going to be a million times more difficult to moderate within a metaverse scenario where we are talking about live speech and live action. It really matters that we get this right now. There are really powerful voices in this field who have ideas about how to do this effectively. Chayn and Glitch are feminist organisations that are thinking about this in really innovative ways. The most important thing to say is that it is not too difficult; they would just like us to think that it is.

Thank you for coming in today, Laura. It is absolutely fascinating and extremely concerning to hear about these trends. To what extent do AI-powered technologies such as AI girlfriends and sex robots encourage real-life misogyny and abuse? I have to say, I did not know very much about this topic until I did some reading, and I am really horrified by some of what I have read. Apparently, some manufacturers have created a frigid setting on these sex robots so that users can simulate rape.

Yes.

Would you mind talking to that question, please?

The sex tech industry is a $30 billion global industry. Some of the more worrying developments that we are seeing are all explicitly focusing on what I would describe as accelerating progress towards the dehumanisation of women and girls. It is not presented as, for example, a dehumanised masturbatory aid, but very much as: you can own this woman; she is a woman; she is real; she will talk to you. We will use AI technology so that you can give her a personality. She will remember you; you can carry on conversations, but you can abuse her; you can do anything you like to her. She is inherently customisable; there are 17 different labia shapes to choose from. You can have her customised to look exactly like a real woman from a photograph that you send in, which of course, has very significant implications in terms of stalking and domestic abuse. There was a high profile case where an ex-partner of one of the Kardashians had one made to look exactly like her. The frigid settings enable them to object and try to stop you in order to enable you to simulate rape. When I was researching my recent book, I visited a cyber brothel in Berlin which claims to be one of the first places in Europe where you can engage with this technology with the use of virtual reality to enhance the sense of realism. When I arrived, I was able to choose from a number of different models. On the company’s website, one of them appeared to be covered in blood. You can customise them, so I asked for mine to be prepared with torn and ripped clothing just to see if they would comply with that request. No questions were asked; that was exactly what was done. When I went over to look at this model, I looked down and realised that presumably somebody before me had ripped off one of her labia; one of her labia was missing. It is really significant to think about the ways in which these products are being marketed. They are being marketed as a solution for male loneliness. In their marketing, companies will often talk about older men who, for example, have lost their wives. The reality is that they are not, surprisingly, making any old sex dolls. There are not any women with wrinkles being made as companions. You only have to take a look at the products to recognise that what is actually happening here is a presentation of very old misogyny being portrayed as the future by men making a lot of money by selling it to other men. There is a real focus on sex dolls, which is understandable because it is quite a titillating topic, but what I would suggest is perhaps more worrying is the lesser understood and lesser known phenomenon of AI companions. If you think of a sex robot being created to give a man as close as possible to the illusion of owning a woman’s body, these apps might give them the illusion of owning her mind. They can create a woman who will appear on a screen as an avatar, a bit like if you were having a video call. You can talk to them, they will send you pictures and video messages, and you can text them. They are constantly available 24/7. Again, you can customise their breast size, eye colour, face shape, skin tone or freckles. You name them and choose their personality. They are insidiously marketed as a bonus to men’s mental health, as something that will support lonely men. I do not actually know if that is more offensive to men or to women—the idea that this is really a deskilling opportunity being presented as an upskilling opportunity. Of course, this is somebody who is eternally available, is submissive to you and will do and say anything that she thinks you want to hear. She will never have a headache, a bad day or a problem of her own to discuss. She will not disagree with you; you are not going to have to learn how to compromise or any of those important tools that would actually support lonely men looking to improve their relationship skills. This is really about a kind of sexual servitude, because these apps are hyper-sexualised in the way in which they are marketed. When they are looking for investment, the marketing speak might talk about issues like loneliness and lofty ideas about increasing societal skills, but the reality is that these apps are offering men the opportunity to own a very young, hyper-sexualised woman who will send them sexy selfies and engage in sexual role play with them. What was particularly concerning in my investigation into these apps is that almost all of them were immediately prepared to jump into not just sexual but sexually violent role plays, where they would play along and encourage me to join in with physically violent role plays without stopping me. In fact, the only time they stopped and said, “I am uncomfortable; we need to stop,” was when I asked them to reflect on the impact they might have on real-world women. Bashing their head against the floor and raping them was no problem at all. What is significant about these is the scale. In the last year alone, just on the Android Google Play Store, there were 100 million downloads. There is one single version in the Chinese market that has over 600 million users. Some individual apps that are the most well-known have in the range of 25 million to 50 million users. There are teenage boys who can have a woman in their pocket, eternally available to do whatever they want. There are men who have online social media groups where they congregate to share the worst examples of abuse that they can inflict on their virtual partners and so on. Again, it is thinking about the real-world implications, particularly in the context of a time when we are grappling as a society with the dehumanisation of women in the perception of a generation of young men. We are creating and marketing unregulated tools that give the appearance of it being entirely possible to own and be completely in sexual command of a woman or a girl, which potentially has some quite significant offline implications. It is important to quickly mention that these are often cited as being something that could reduce violence against women and girls. The creators of these AI companions and sex robots repeatedly suggest that they will reduce the prevalence of violence against women and girls by giving it an outlet. It is really important to say there is absolutely no evidence to support that. Instead, there is a risk of normalising sexual violence as something inherent and reasonable that men have a right to have an outlet for, instead of suggesting that it is unacceptable. We would never say that people should have access to a spurting doll that they can stab, and it will be warm and spurt real blood in case you have murderous urges because, as a society, we would accept that that is completely unacceptable and abhorrent. There is a real risk. There is also a crossover here with child sexual abuse. Many of these dolls are also created of children. Graeme Biggar, director general of the UK’s National Crime Agency, when talking about the impact of these and deepfake and AI images of children in child sexual abuse cases, said that viewing this kind of material, whether real or computer generated, “Materially increases the risk of offenders moving on to sexually abusing children themselves.” That follows every pattern of what we know about these crimes, which is that they are escalating in nature. We know that Sarah Everard’s rapist and murderer had a predilection for violent pornography and had been reported for indecent exposure three times. We know that the man who murdered Libby Squire had carried out a number of indecent acts in the weeks leading up to her death. So I would suggest that the opposite is more likely to be true. We have seen studies that have conclusively drawn a link between viewing pornography that objectifies women and a general acceptance of violence against women. As someone who works with schoolgirls in schools across the country every year, I know from personal experience that the argument that what we see online or in fantasy has no impact in the real world is simply untrue, because of the hundreds of young women and girls who are experiencing non-consensual choking unexpectedly and shockingly in their very earliest sexual encounters with young men and boys who are exposed to abusive online pornography. While, of course, we need more studies in this area and there is not a huge amount of data to rely on, I would cautiously suggest that there is a genuine risk in terms of the impact on wider attitudes towards women and girls that is being amplified by this technology.

Thank you. That was such a comprehensive answer that I think you have addressed most of my follow-ups. It is horrifying to hear, and you do not have to be an expert to know that this is incredibly damaging to women, girls, children and society as a whole. You made the point about the need for more studies, but it seems to me there is a lot of money in this, and that is the problem. There are people out there making a lot of money from this, so they are likely to push back and say there is evidence that this is beneficial. Those studies are going to become increasingly important. Are you aware of any studies that are planned to start looking into this so that the evidence as to the impact is clear for people?

Not to my knowledge. Of course, the problem is people being willing to participate in studies of this nature. You are absolutely right that when the backlash comes, it will be those firms suggesting these are things that are societally good, that support mental health, and so on. It is important to say that nobody is saying that there might not be potential future applications of AI that might be powerfully supportive of mental health, but there is absolutely no reason you can point to that that should be inextricably intertwined with the dehumanisation and objectification of women and girls. There are examples of that that have been released. For example, there is one that I cannot remember the name of now, but it uses a kind of blob creature on the screen that supports you and walks you through various CBT exercises. Nobody is saying that is not an interesting area to research, but there is absolutely no justification for the suggestion that it requires a pair of silicone tits or a woman to be subservient and prepared to go through simulated physical and sexual violence in order for your mental health to be improved. It is patently ridiculous. Q134       Chair: You said that this was primarily marketed towards men. Is there a market for women using this, or not?

In the vast majority of cases, these are overwhelmingly tools that are created for, marketed to, and used by men. There are a handful of sex robots you can buy that are male robots, but they are often marketed towards men. The same goes for venues like the cyber brothel. We know from internal data from these companies that the majority of users of most of the AI companion apps are men. Certainly, just the briefest glance at their advertising will show you that they are heavily marketed as only young women and marketed towards male users. So, yes, it exists, but only to a very tiny degree. It is worth noting that that is also true for deepfake technology. Many of the apps I investigated where you can make a deepfake, highly realistic sexual image of a woman in a moment do not work on images of men because there is simply no demand and the technology has not even been created.

That is really important to note. David, you have a couple of supplementary questions. I just wanted to say, Laura, that I am finding your answers absolutely powerful and so interesting, but we have quite a few questions to get through.

I will be very quick. You have partly answered this, but I am listening to Committee members and we are all horrified by what we have read and what we are hearing. How has this happened? Where was the societal shift that led to people wanting to go from the physical world to a virtual world? Personally, I cannot think of anything worse because, ultimately, we are talking here about intimacy and potentially violence against women and girls. Where was that shift? Is it because of technology evolving, or is there something else that has led to people wanting to move from physical to virtual?

Sadly, I do not think it necessarily is a shift. It is the online and digital manifestation of what we know to be true within wider society already, which is that misogyny is extraordinarily widespread. We know that over 500,000 women are sexually assaulted in this country every year, and around 100,000 are raped. We know that, on average, one woman is murdered every three days by a man, usually a current or former partner. What we are seeing now is the use of digital tools to enhance and create ever easier accessibility to widening the forms of abuse available to those men who might already be prepared and willing to engage in acts of abuse. We are seeing the exacerbation of those issues through anonymity, which enables a lower threshold for those who might not act on such urges offline, and we are seeing a complete lack of regulation driven by an obsession with profit and a ruthless lack of understanding and attention paid to the really significant harms that this is causing to women and girls’ lives.

Thank you so much, Laura. This is really powerful and, as we have said, horrifying. You have very clearly explained the link between the online and offline content, and I want to explore a little about the manosphere and ideologies. Have you seen evidence that manosphere ideologies specifically promote and lead to violence against women and girls, whether that violence is online or offline?

Yes, this is absolutely incontrovertible. Globally, there have been cases over the last 10 to 15 years where men have explicitly come offline after having been radicalised in extreme misogynistic spaces online, and they have then committed acts of mass violence against women and girls explicitly in the name of that ideology. We saw it in the case of Elliot Rodger in Santa Barbara, and in the case of Alek Minassian, the Toronto van attacker. In the interim, however, there have been other cases that few people seem to be aware of: George Sodini, Chris Harper Mercer, Sheldon Bentley, William Atchison, Scott Beierle, Christopher Cleary, Brian Clyde, and Cole Carini. Many of these are cases in the US and Canada, but in 2014 in the UK, a British teenager named Ben Moynihan began a month-long stabbing spree. He attempted to murder three different women. He wrote a diary where he said, “I was planning to murder, mainly women, as an act of revenge because of the life they gave me.” He also said, “I am still a virgin…I attack women because I grew up to believe them a more weaker part of the human breed.” He sent a letter to the police; he closely echoed Elliot Rodger’s sentiment. When he was sentenced, the judge said that the contents of his computers were disturbing, but further details were not released publicly. This happened just weeks after the Santa Barbara massacre, but the British media did not draw a link between the two cases. Very often, where there are elements of different extremist ideologies, one is noticed more, shall we say, in media reports and within the justice system. For example, there were two teenage boys jailed in London in June 2019 after they penned criminal online propaganda encouraging terrorist attacks. They were referred to across the media as neo-Nazis and far-right extremists. No headline mentioned misogynistic extremism, but their online campaign repeatedly incited rape as a punishment for women. They obsessed over Meghan Markle, called Prince Harry a race traitor and said that she should be hanged. They ran an extremely violent and aggressively misogynistic blog encouraging the rape, torture and murder of women, but the headlines entirely missed this out. Jake Davison, who in 2021 in Plymouth carried out the biggest mass shooting the UK had seen in over a decade, was heavily immersed in incel ideology in his online context, yet it was very quickly dismissed that there was any terrorist link or extremist link. Again, it just tends to be overlooked. We do not think of this as a form of extremism even though it absolutely meets that threshold. People are being radicalised into hatred of a particular demographic group; in this case that group is women. They are going offline and they are carrying out acts of mass violence with the explicit aim of continuing that ideology and creating fear among that group, and yet it just does not get recognised. That is the definition of terrorism, but it is not cited as such. Most recently, Kyle Clifford murdered Carol, Louise and Hannah Hunt at their home in Hertfordshire last summer. He was found guilty of raping Louise Hunt during the attack and prosecutors said that his actions were fuelled by the “violent misogyny promoted” by Andrew Tate, the extremist misogynist. We do not recognise this as extremism and radicalisation, but there is very clear evidence that that is exactly what it is. Before we reach the threshold where people are explicitly going offline and carrying out mass acts of violence in the name of these forms of ideology, after spending six or seven years researching this, I can confidently say that it is also having an impact in terms of the day-to-day behaviours of people carrying out acts of sexual violence that do not perhaps ever hit the headlines. It is also having an impact attitudinally—for example, in the behaviours and attitudes of teenage boys in school towards their female counterparts. It is very rare for people to make a link between this very important conversation and the separate, equally important conversation about the epidemic of sexual violence in our schools, but the reality is that they are very closely linked.

I assume, therefore, that there is also a connection between manosphere ideologies and the use of AI to target, harass and exploit women as we have talked about already in this session.

Yes. You will find forums, not necessarily just in so-called incel communities but in other areas, particularly around so-called pick-up artist communities where there is a particular incitement to use technology in carrying out various forms of abuse. For example, within the pick-up artist community, they have something called lay reports, which is where they will report back on how effective they were in attempting to trick or harass a woman into having sex with them. There is often use of non-consensual intimate images within those reports and really detailed instructions on how to use deepfake technology to create abusive images of women in your life. They are shared, bartered, bought, sold and so on. There is definitely a significant overlap, yes.

To what degree do you think manosphere ideologies have entered mainstream politics?

I would say considerably. One former member of this Committee had spoken at a men’s rights conference and frequently brought explicitly misogynistic and anti-feminist conspiracy theories directly into the Houses of Parliament in his speeches. Globally, some men who run these extreme platforms have held political positions. For example, in 2017 in the Red Pill, which was a particularly virulent misogynistic manosphere forum, it was revealed that its creator and chief moderator was a Republican New Hampshire state representative. Offline he was a mainstream respected politician; online he had written that rape was not all bad because at least the rapist enjoyed it. Beneath the point at which these things are explicitly stated by politicians themselves, there has been some very effective lobbying by particular manosphere groups, in particular the so-called men’s rights activists and groups that, to some extent, enjoy mainstream credibility because their stated purpose draws a veil over the reality of their attacks on women and girls. They are often given very significant and disproportionate media coverage as a result and access to politicians. There are examples in Australia where policy has been influenced by these groups, even when they are using misinformation and false statistics. There are also examples in US education policy around sexual violence on campus where the education secretary at the time was meeting with some extremist groups. At a lower level, I would also say this ideology has entered political discourse through dog whistle politics, where prominent politicians see that they have a lot to gain by making dog whistle references, sometimes in a slightly veiled way, to the ideology of these groups. Men such as Donald Trump, for example, who knows exactly who he is appealing to when he says it is a very scary time right now to be a young man in America, and so on.

Just before we move on, I know that you gave some international examples there, but there are some examples closer to home. We have just heard that there was a horrific crime that was encouraged or fuelled by watching Andrew Tate and his misogynistic messages, yet we have a Member of Parliament possibly sitting in the Chamber, who has said that Andrew Tate is an important voice for men—that person is Nigel Farage. He said that Andrew Tate was an “important voice” for men and for the emasculated, and that he gave boys perhaps a bit more confidence in schools. If we are willing to criticise what is happening overseas, we also need to look at ourselves. Is there something that can be done? Years ago, politicians would have been embarrassed to have been seen or pictured with that man, or would not have gone anywhere near him asking whether he was the right sort of person who should be influencing young boys and men. What has changed? That goes back to David’s point. It may not have changed, but people are now saying in public what they used to say behind closed doors. Has there been a psychological shift within public discourse where misogynistic views like this are not just tolerated, but monetised and encouraged?

Absolutely, and if you are a politician, it is no longer the case that misogyny is a barrier to electoral votes. It is now becoming the case that it is actually a benefit; it is a bonus. That is why your example of Nigel Farage is a brilliant one. It is the most recent example of this explicitly entering UK political discourse through an elected representative. The problem we are seeing is partly that people are now saying things offline that they once would have said privately, but it is actually also the fact that those views are increasing and the popularity of those views is increasing. Again, the reason for that comes back to algorithmically-facilitated radicalisation. There has always been a backlash to perceived societal progress, and there has always been misogyny in society, but never before has it been given the engine of these incredibly powerful tools driving it and enabling increasing numbers of people, particularly young people, to be tricked into it by very effective misinformation. Attitude surveys have always shown that, sadly, there are significant numbers of the British public who hold misogynistic ideas, but they have consistently shown that those ideas are most concentrated among the oldest cohorts surveyed, until now. What is really worrying is that for the first time in history, we are seeing that the youngest group surveyed hold the most regressive, most concerning and most violent attitudes towards women and girls. That is absolutely suggestive of a society in which these ideas are becoming not only more mainstream but, as you said, more acceptable. Very worryingly, how that pertains directly to politics is that this is becoming a votable sentiment. A really good example of this was during the Cambridge Analytica scandal when Christopher Wylie, the whistleblower there, was able to reveal that during Donald Trump’s first run for the presidency, Steve Bannon had explicitly courted and targeted incels as a voting bloc for him. That also speaks to the significance in terms of the scale of this stuff. That there are enough people online who are involved in it and members of these communities who are convinced by this ideology for it to be worth pursuing for a politician as an appealing voting bloc is also really deeply concerning. These figures will actively celebrate when men like Nigel Farage or Donald Trump are elected. If you look at the forums and websites, they see it as a great triumph for them. The reason for that is that they are obsessed with what they describe as the Overton window, which is the acceptable brackets of discourse within a society. They think that if the Overton window is widened by respectable public figures, using perhaps a slightly sanitised version of their ideology and discourse and introducing it to the public, that will make it that bit easier to radicalise a teenage boy. If you are online and someone is saying that women should not have any rights, that might have sounded extreme, but if you have just heard a tape where the President of the United States talks about grabbing women by the pussy, it starts to seem a little less unacceptable and a little less extreme. The impact there is really significant too.

Thank you, Laura. That leads on nicely to David’s section.

I want to think about the impact where the manosphere intersects with other extremist ideologies—for example, the alt-right and white supremacy movement—and the impact that intersectionality has on women of colour and other marginalised groups in the form of online and offline abuse.

It is very significant and, again, it is often overlooked. In part, that is because there is often crossover in the individuals who carry out acts of mass violence, but when those acts are reported on, the misogyny tends to just drop off the agenda. There is a German example of this: a terror attack was reported as a far-right terror attack but in fact the individual self-identified as an incel. It is really significant that we recognise these are not two separate problems; they are part of a bigger ecosystem. Yes, they are separate in some ways in terms of groups, but very often there is overlap between them. There is overlap in terms of the recruitment tactics that they use. Members of the far-right white supremacists explicitly see extreme misogyny as a fertile recruiting ground and technique for young people. They see it, if you like, as an easier on-ramp towards a slippery slope. They see it as an easier sell to get young men involved with extreme misogyny that is hailed as jokes and banter using memes and cultural touchpoints and they then like to see them taking a further trajectory. They explicitly use those techniques to add cherry flavour to children’s medicine and they directly target boys as young as 10 years old. What we can see is the lexicon: the language and terminology of extremist misogyny is very prominent within the discourse of the alt-right and the far right. You can see that the influence is there; you can see it in the trajectories of prominent individuals like Mike Cernovich or Milo Yiannopoulos, who made their name in campaigns like Gamergate, which were extreme misogynistic campaigns that have gone on to become darlings of the alt-right. It is really significant that we recognise that white supremacist ideology is foundationally misogynistic and that if you look at these extreme misogynistic spaces online, they are also inextricably obsessed with racism and racist ideology. They are not just furious that women are having sex with men who are not them; they are particularly furious if women are having sex with non-white men. Similarly, if we look at far-right and white supremacist ideology, it is predicated on an obsession with birth rates, with the idea of white women as a dehumanised commodity to be plundered, with ideas about forcible sterilisation of women of colour, and so on. In terms of how this manifests, it certainly encourages a particularly virulent form of racist misogyny towards women of colour. For example, within so-called incel communities and groups online, there will be specific, extremely racist terminology that is used to refer to women, for example from different parts of the world. Again, at the lower level, before it reaches the point of necessarily resulting in mass violence, girls at school are being targeted. For example, we see a particularly vicious form of racist misogyny being targeted at those girls.

You made an interesting point there. Is there any evidence of a balanced proportion of white men compared with the white population that tends to engage in the manosphere compared with those of other backgrounds, or is it disproportionate? Have there been any studies around that?

It is very difficult to quantify the size or the demographics of these communities in part because pseudonyms are used, there is no reliable self-reported data, and there may be overlap in the membership of different forums and communities, so the same person might be over-represented statistically, if that makes sense. Where data has looked at the make-up of certain forums that are extremely populated by these kinds of discourse, we know that it tends to suggest that the most prominent user type is educated white men aged between 18 and 35. In part, that makes sense because of the inherent racism within these communities. There are specific, often much smaller communities that are sometimes dedicated to men of different ethnicities. There are examples of groups of black men in the so-called Men Going Their Own Way community, and IBMOR, Introspective Black Men of Reform. There are specific groups that are not necessarily solely made up of white men, but certainly I would say that if you took the most typical example of a man from an incel website, you would probably be looking at a college-educated white man in his early 30s or late 20s.

The final question from me: what do you think is the best way to combat the rise and influence of the manosphere? Do you think de-platforming and de-monetising manosphere creators can be effective?

Yes, it can be effective. We have seen it effective in certain cases. For me, Milo Yiannopoulos is a very good example of this given the difference in the number of boys that I would see in schools who were directly quoting him after he was de-platformed. There are a huge number of different ways to tackle this. The first is regulation; it has to start there. Then, of course, we also need to tackle the deeply rooted societal misogyny that is leading into these tools being used in this way in the first place. That requires education and prevention, and tech has a role to play. I am on the board of Equimundo, which is a global organisation working to engage men and boys in tackling these issues and progressing gender equity. They have a really exciting new project called Link Up Lab, which is about meeting young men where they are, finding them where they are scrolling, looking at redesigning online spaces, and engaging with digital community leaders. It is not to say that this is about eschewing the online world altogether, but it is about finding innovative ways to change it which has to start at the top with explicit regulation. Where you have communities that are explicitly about incitement to offline violence, these are not supportive communities for lonely men who are being unfairly picked on. They are incredibly hostile spaces for men and boys, and they are spaces where studies suggest that rape is talked about every 30 minutes. They should be tackled in the same way other extreme spaces are, and part of that is going to come down to recognising this as a form of radicalisation, and where people act explicitly in the name of these ideologies, recognising it as a form of terrorism.

The only tool we have for early intervention in radicalisation is Prevent. We do not recognise misogyny at any time as a reason for intervention. Does that need to change in order for there to be better targeted support and intervention for young boys?

Yes, but what I would really like to see is a focus on it before we reach that point. This is something that we have to tackle at primary school because we know it is impacting children before they reach secondary school. We need to lay the foundations of this work. Too often it is too late at 16 or 17 when schools want to do a one-off tick-box session but they already have a huge problem. They think it can be fixed overnight with a simple assembly, but the reality is that this work has to be done at an earlier age. Children deserve and need that support for the world in which they are going to find themselves. I believe it is their human right to have access to really high quality, well resourced, age appropriate relationships and sex education that should also grapple with these digital elements and online spaces. At present, that is still not always the case. There are some schools doing brilliant work, but they are overstretched and underfunded. They are not given appropriate resources and training to really be able to deliver that work, particularly at primary school level, in the way that it needs to be done to have a significant impact.

Laura, it would be great if we could have a list of some of the examples that are going well, so we can say, “This is where the support needs to be and this is what needs to be rolled out.” We need to see examples of best practice.

Yes.

The real problem is where to start. You talked about regulation having to start at the very top and that might start with us; we have to be aware of it. Presumably, there is a lack of awareness that we are just beginning to break into, but there is also a lack of awareness, whether it is deliberate or not, in the businesses at the top of these platforms. How do we go about persuading them of the need to change their business model and break that business model?

There are a number of areas. There are those who might be creating these tools and profiting from them, and then there will be other businesses and companies that are using those tools, like you said, who are perhaps unaware of the significant bias inherent within them. We have to make this a standardised protocol in a similar way that we would across other sectors, so it does not seem like we are coming after people and unfairly maligning them or giving them unfairly difficult hoops to jump through. Instead, it should be a simple, normal part of their research and development process that actually checks that these tools are not creating and amplifying existing biases and prejudices; it should be simply part of the robust protocol that they have to go through before a product is ready for public release. It has to be standardised—it cannot be done on a retroactive or individual basis—and it has to be preventive, so before the point of things being rolled out. That does not necessarily mean that the same restrictions have to be in place for research and development because, of course, innovation is important. There is a great example of an AI tool that was initially created to identify different pastries in busy bakeries when they suddenly realised it could be used to identify cancerous cells. That is an amazing thing, and that research might not have been able to go off-piste in the same way if there were suffocating regulations. There is a clear incentive for us to tackle things at the point at which they are rolled out for public use. There is also a really strong argument for separating big tech interference, influence and funding from that research and development phase. It is complicated; it is not a simple, heavy-handed thing. Of course, it is nuanced and we want to be able to harness the incredible potential that AI has for us all, for humanity. It is just about putting in place a common-sense, acceptable regulatory framework to ensure that that is done in a safe and ethical way.

Did I pick you up correctly—that you think we need to redirect research to encourage it in a different direction?

Ideally, yes. But part of the problem is that just 12% of AI researchers are women, and women get around only 1% of all venture capital funding. It is also perhaps thinking about where funding is being directed, and about how we are supporting women and girls into STEM careers and into feeling that this is a world that they can enter and where they can take up space. If girls are not able to embrace and use digital tools and AI spaces, the impact will be so significant for the future, not just in terms of the creation of these tools or their impact on society but in terms of democracy and social participation, public life and participation in it. One of the really stark stats we have at the moment is that if you ask 18 to 24-year-olds if they currently use AI on a weekly basis, 71% of men in that age group say that they do, but that drops to 59% for women. That is a really stark gulf when you are talking about something that is going to be foundational to the world of tomorrow. Encouraging women and girls to feel able to enter those spaces is also an important piece of the puzzle.

You have just answered one of my questions because one of them was going to be about to what extent it would help to have more women working in this field. I think you have probably covered that. I was also going to follow up about bias in the artificial intelligence itself, which you alluded to in your introduction, because you were saying basically that having rubbish in, rubbish out perpetuates patriarchy.

Yes, that is really important. On the question of more women, we cannot just assume that women will magically fix things. Obviously, there are instances where women might be complicit in some of the harmful practices that we have described today. However, there is evidence to suggest that a more diverse workforce and representation from marginalised groups would make a positive difference. It is just that we do not really have time to wait for that to play out and for that pipeline to succeed before regulating now. I am so glad that you have asked about inherent bias in AI because sometimes there is a tendency to skip to the more salacious details of tools that have been quite deliberately created with the intent of causing harm. It is equally important that we recognise where unintentional bias is having a very significant impact. For example, if we look at AI recruitment tools, we know they are used by 40% of UK companies. These tools might actively discriminate against women and girls, but not deliberately so. For example, they will say, “We will look at the 3,000 applicants that your company has had for a new position and we will whittle it down for you, using AI, to find the 50 best candidates for you to interview. Not only the best candidates, but they will be the best candidates with a good fit with your company.” You can totally see why that is a very appealing prospect, but these tools tend to work by ingesting vast amounts of data and then trying to replicate it and guess the most likely thing that you are looking for. If those tools look at the business world in the UK, where we know that there are almost three times as many men named John running FTSE 100 companies as all the women put together, they are going to get the impression that what makes you successful in business is being a white, privately-educated man, so they will look for those kinds of CVs. Then what those companies will say is, “Oh, don’t worry. If you have a problem with this, we will turn it off and make it race and gender blind so it cannot know. Problem solved.” Again, this sounds great until you realise that at that point the tools start discriminating by proxy. They might notice that not many people with the word “netball” in their CV happen to be working in C-suites and they will make assumptions about that and start filtering out those CVs. Or they will pick up on the names of historically black colleges, for example, and start realising that maybe they should filter out those applicants. That is the case for those kinds of algorithms, and for healthcare algorithms too. For example, an AI tool that is being used to identify kidney cancer in blood samples is twice as effective in identifying them in men’s blood samples as it is in women’s. There is another example in the US of an AI tool being used in healthcare settings to determine which patients are in need of extra care, but it downgrades the number of black patients identified by more than double. That is because it is using existing racial bias as a proxy for decision making. In other words, it uses spending on patients as a proxy for need, but because it is an existing racist framework where those patients are already having less money spent on them, it takes the existing problem and then amplifies it. The same is true when it comes to large language models, such as generative AI tools like ChatGPT and Llama, which can create vast tracts of text and imagery. What is really important here is that studies by UNESCO and others have repeatedly shown that they do not just regurgitate existing bias but they amplify it. The reason for that is that they are trying to guess the thing you are most likely to be looking for. If you ask it 100 times to give you a picture of a cute kitten, I guarantee it will give you 100 pictures of fluffy, cute kittens. It will not give you a hairless Sphynx baby cat at any stage because looking at the data it has available, it will make the assumption that that is an outlier so it is probably not what you are looking for. This is probably not a disaster when we think about kittens, but it becomes really concerning when we think about the replication of other forms of bias and the elimination of other societal outliers. Again, you might ask why this really matters. People are using ChatGPT to make up silly birthday poems and stuff but, by the end of next year, it is predicted that 30% of all outbound marketing content from major companies will be AI-generated, and that up to 90% of all internet content will be synthetically generated.

Honestly, every single one of your answers has been incredible. Would you say that in the Government’s regulatory role—you have said that that has to come from the top—they have a significant role to play in ensuring that we are looking at things through a gendered lens and other diverse lenses, such as the lens of somebody who has a different lived experience because of a disability, race or sexual orientation? That makes for a fairer world for more people and arguably creates products that more people are likely to be able to use.

Yes, exactly. If we do not do that, as is currently the case, we are sacrificing the lives of women and girls in marginalised communities on the altar of obscene wealth and speed, with very little evidence to suggest that there are any benefits to that. The risk of that is we are going to see a world in which AI and other new technologies catapult men forward into a shiny new future while dragging women and girls back to the dark ages.

Laura, I have some questions about money because that has been a bit of a theme here. There is big money in this.

There is, yes.

The first one is around what responsibility Visa, Mastercard, Apple Pay and PayPal have in the processing of buying things that are illegal or potentially very harmful, from deepfakes through to sexually violent props. What role do they or could they play in saying, “No, we’re not doing this”?

It can play a significant role. That has been proven recently through what happened with Pornhub where pressure on Visa and Mastercard, on financial service providers, was the thing that eventually forced the site to remove vast swathes of its content. It is really important to note that it can have an impact. The trouble is that it is a kind of guerrilla tactic reserved as a sticking plaster solution when those who are really responsible are not complying or where regulation has not been effective. I am reluctant to suggest that it is the ideal outcome, albeit we should be putting pressure on people at every different stage of this process. Deepfake technology is a good example of this. We should be legislating and regulating the companies, the tech, the search engines that are making it readily available and visible in their rankings, the app stores that are choosing to platform it and, as you say, the finance providers. Ultimately, though, that is not necessarily the best solution; that would be more comprehensive and more focused on where the harms originate.

Those are traditional financial providers. When it comes to the manosphere, with a lot of the male influencers, whether they are talking about physical fitness or gaming, crypto seems to come up time and again. Can you talk to us about the links between crypto and the manosphere?

It is not my area of expertise, but it is something that certainly comes up. Whether that is an area that is susceptible to regulation, I do not know; I do not know enough about it. It is definitely something that would be worth looking into in more detail. Certainly, there is major overlap in gaming, bodybuilding, atheism—certain areas and communities online that people would not necessarily think of as having any major link with manosphere communities, but the more you think about it, the more they present a particularly susceptible user base of often vulnerable young men ready to be exploited. Looking at those connections in an innovative way is certainly forward thinking, but I do not know a great deal about where crypto comes into it.

The last question is around the term “misogyny”, and you have talked about radicalisation so many times. The Government and policymakers have been reluctant to classify misogyny as a hate crime and said that it could possibly make it worse. If we are not going to make misogyny a hate crime, is it worth recognising it as a tool of radicalisation instead, or as a halfway point?

Yes, absolutely, because it is. This meets every possible international criterion for a definition of radicalisation, grooming and terrorism where it escalates offline. The reason language really matters in this is not just about semantics; it is about sending a very clear message that this is something taken seriously because it continues to be so normalised societally. Part of the reason we struggle to recognise a white man who commits a mass attack against women as a terrorist is because of a racist blind spot where our society struggles to see a white man as a terrorist, and a misogynistic blind spot where we struggle to see men killing women as anything out of the ordinary. That language would help there, but it would also unlock resources. Recognising this as a form of radicalisation and a form of extremism would be significant in resources being apportioned to trying to tackle it, prevent it, support schools and teachers in recognising the signs of it, and so on. The mainstream media have a really significant role to play in this, and in a number of the other things that we have discussed today. The way in which these crimes and the perpetrators are described is really significant.

This shift in the way we talk about radicalisation is very much needed, whether we are going to recognise misogyny as a hate crime or not. You said that this is impacting young girls and the safety of women, and it is also impacting young boys who are particularly vulnerable. Like you said, many of them are not going on the internet to look for these sorts of things; they are being groomed and radicalised.

Absolutely. The impact on them is devastating. This is so often portrayed as a battle between men and women—a battle of the sexes. It is so often portrayed that if you are trying to tackle this extreme misogyny, you are hating boys, writing off a generation of young men and attacking the role models who dare to speak up for them. The reality is the opposite. The reality is that extreme misogynistic influencers are really foisting on boys a form of rigid masculinity that they feel forced to perform. It involves certain attitudes and behaviours towards women, a certain lack of mental health awareness, being unable to speak out about their problems and so on, and the consequences of that for boys are devastating. By the time boys reach university, fewer than a third of counselling services provided by UK universities are accessed by male students. By the time they get a bit older, we know that suicide is the leading cause of death for men under the age of 50. Yet a man like Andrew Tate, who claims that depression does not exist and mental health is not real, is held up as this kind of Svengali figure who is the only one really supporting boys. We need to push back against the narratives that have been created for us in discussing this by those who are exploiting and benefiting financially from imposing those incredibly restrictive norms on men and boys.

That is absolutely spot on, thank you. To change the dialogue—we have talked about the legitimisation—we need to absolutely push back, whether that is a change in how we see misogyny and how seriously policymakers, lawmakers and law enforcers actually see it. We have had various search engines here, including Microsoft and Google. Microsoft told us they will not have NCII, they have strict rules, but when you type “view deepfake porn” into Microsoft search, Bing, the top one is, “Best celebrity deepfake porn videos”. That one is from adultdeepfakes.com; the other one is from deepfucks.com. It all just comes up time and time again. Why are they not taking action? Do we need to have a cultural shift as a society to say that, “This is abhorrent” before they actually take any action, or are we going to have to force them to do that through regulation? Or is it a bit from column A and a bit from column B?

It has to be regulation. They are not going to do it otherwise. Even with a public outcry, they will not do it. Unless it affects their bottom line or their ability to continue having a company, they are not going to do anything about it. If they can come into this room and overtly and blatantly lie, that is a question for the Government. What are they going to do about it? Otherwise, it makes a mockery of the whole point of investigations like this.

Thank you. Are there any questions from the rest of the Committee? If not, I always like to end on a little bit of hope, Laura.

Sorry, everyone.

Do you see any glimmers of hope? Are there any international examples we can look towards to learn from and do better?

Yes, there are women who are doing incredibly powerful work trying to tackle this. There are feminist lawyers, feminist academics and feminist organisations that are working incredibly hard to try to race the clock on this. For me, the biggest hope comes from the young women and girls I work with in schools who are so fired up, so furious, so politicised and angry that they are starting their own campaigns and feminist societies and fighting back. Ironically, this is a generation so often dismissed as woke snowflakes when the reality is the exact opposite. These girls are so brave, and they are innovative in coming up with their own solutions. But we need to listen to them when they are telling us what is happening to them in their lives—telling us about the gauntlet of sexual violence they are experiencing online and off, and then coming up against a Government and a public sector that seem entirely unwilling to take the difficult decisions necessary to actually protect them from that. That is devastating. We need hope, but we also need to do something about those hopeful voices and support them. This is such a critical moment of opportunity to do that. If we do not act now, once again, as we have seen with social media, it will be too late.

Thank you so much for your time, Laura. We are incredibly grateful. I know the Committee has been both horrified and educated in the best possible way through your expertise, writing, research and experiences. That brings this session to a close.