ディープフェイクの被害者となった英国議員 Jess Asato、xAI を提訴へ

Your workflows span systems. Your data should follow. Dun & Bradstreet's context layer delivers verified data across your AI workflows so your AI can reason and you can take action. Give your AI the context that's missing. Visit dnb.com slash trustyourai. AI can make businesses faster. It can also make mistakes, expose data, and execute the wrong instructions faster. ThreatLocker helps organizations set boundaries AI cannot ignore by controlling what applications can run, what they can reach, and how they interact with sensitive resources. Innovation needs more than good intentions. It needs limits. Control AI with zero trust at brettlocker.com slash wired. That's T-H-R-E-A-T-L-O-C-K-E-R dot com slash wired. From Wired, this is The Big Interview. I'm Deputy Editor Rosie Swash, sitting in this week for Katie Drummond. My guest today, Jess Asato. is a British Labour MP who, after years of campaigning to protect women and children from online image abuse,

became a target of AI-generated abuse herself. Now she's taking on the tech industry, Elon Musk specifically, and his company XAI, to try and establish legal precedent that forces AI companies to implement safety-by-design guardrails and hold them liable for how their systems function. Hi, Jess. Thanks for being here with us today. I just want to have a couple of warm-up questions before we get into it. So if you wouldn't mind telling me a bit about your background, how you ended up becoming MP for Lower Stuffed? Sure. So I grew up in East Anglia in Norfolk and Suffolk and not in a very political family particularly, but I got the politics bug in 1997 when Tony Blair was talking about education, education, education. and I felt that we could be doing a lot better in our education system, having experienced it and joined the Labour Party,

got involved with my local groups. The last 10 years before Parliament, I worked for the largest children's charity in the UK on their policy and public affairs, so trying to change legislation on behalf of the most vulnerable children in the country. And then six years prior to that, I worked for a domestic abuse charity called Safe Lives, where we did huge amounts of work to try and improve the situation for survivors of domestic abuse, particularly through the Domestic Abuse Act. And those twin issues of women and children have become my focus in Parliament. And I've been very clear that my reason for being here, you know, my mission really is to be a voice for women and children. Well, that's a good entry point for us to start talking about this year. So I wanted to take you back to the beginning of this year, January 2026. There was widespread reporting of Grok XAI's chatbot being used to create sexualised images of children and women and those images being circulated on X.

How aware were you at this point of the phenomenon of deep fakes being used in this way? very aware and actually have been working with survivors who had experienced deep fakes for quite a number of years ever since the advent of ai this has been happening to women and the developing risks of children both being put into child sexual abuse material but also children using nudification tools on other children, thereby criminalizing themselves because these tools are so easy to access, often free. So I was already very alive to the huge harms being created by AI deepfakes of both women and children before Grok ever happened. It was a known harm, which is why many of the tools from other companies had already taken steps to stop them from being used to create this kind of content.

And there are a number of tools where, you know, you type in nudify or take the clothes off this person and the tool will refuse to do that. So into that environment came Grok. Was there a particularly egregious example, an incident prior to Grok when you became aware of deepfakes that really stuck with you? In particular, there is a brilliant campaigner who goes by the name of Jodie. I think it was probably listening to the experiences of Jodie, who was deep faked, and it was such an affront to her that somebody would do this. And then she found out it was actually a friend. And these things destroy not just the individual involved, but often families and everyone around them too. For people who don't know, who exactly was Jodie? Is this a constituent of yours or someone who came forward to you specifically? No, no, no. She's a very well-known campaigner on this issue. That's not her real name.

Before, what happened was that victims had to prove that there was no consent and now it's up to the perpetrators to prove that there was no consent to create these images. I know it sounds obvious, but what exactly is nudification? Well, it is where you ask an AI tool to take the image or video of a real person and using generative AI to make that person either nude or engaging in pornographic acts. This is something that because of the huge development of AI and the real photorealistic nature of it, it has become increasingly hard to see this as a fake image. It can often look incredibly, incredibly real. So what was it particularly about the Grok scandal earlier this year that made you want to raise this in the House of Commons? It was the fact that we already knew that you could stop this from happening by designing the AI tool to reject the ability to do this.

In fact, it actually says that the tool should accept asks for the creation of adult sexual content, even that which has dark or violent themes. And the fact was that this was completely avoidable. unavoidable. And that's why I spoke out, but also because I'd just spent a year and a half campaigning to get a ban on nudification apps. And here was an all-purpose AI tool becoming a nudification app. And it just felt to me as though we were making strides in the UK. And then the world's richest man and his product decided to turn the clock back. So what exactly did you stand up and say? What I did was I talked about the need for the government to accelerate their ban on notification apps, which they did. So the Secretary of State, Liz Kendall, said that she was going to bring that ban into law through the Crime and Policing Bill, which she did. But I also said that it was very important that we went further than that, because of

course grok itself is not a sole purpose notification tool it has a general use and therefore we need to bring all ai chatbots under the auspices of the online safety act and we're now waiting until that legislation comes into effect there's got to be a consultation i understand before it does so technically another grok could happen in in in this period but in the future, there would be much stronger measures that we could take against tools like Grok, creating needification content. Take your curiosity with you on KUOW's new podcast, Pocket Science. In each episode, we catch up with local scientists in the lab and in the field to find out how they're pushing the boundaries of what we know. From the surface of an icy moon of Saturn to the inside of a human

brain on psychedelic drugs. The answers might surprise you. Listen to Pocket Science on the KUOW app or wherever you get podcasts. AI systems don't operate in a vacuum. They need context, and that context comes from data. Choose the wrong data and your AI could deliver something between hallucination and guesswork. Dun & Bradstreet Context Layer, anchored in the Duns number, delivers the context your AI systems need to reason, decide, and act across your workflows. Business identities are resolved, relationships are clear, and continuously updated data keeps those systems grounded so your AI doesn't guess and you can operate with confidence. Give your AI the context it's missing. Learn more at dnb.com slash trustyourai. AI tools can write code, access files, connect to the internet, and execute commands at a speed no human can match. That productivity is powerful, but it also means one bad instruction or compromised tool can quickly become a security incident. ThreatLocker helps organizations use AI without losing control.

Put zero-trust guardrails around tools like ClaudeCode and control exactly what they can execute, access, and connect to. Set clear boundaries around the applications, files, systems, and online destinations your AI tools are permitted to use. Because approving an AI tool should not mean giving it unrestricted access to your environment Control what runs Control what it can do Control what it connects to Learn how ThreatLocker can help you safely embrace AI at ThreatLocker slash Wired That T dot com slash Wired So that's a really positive outcome of your interjection in Parliament. But what else happened as a result of you standing up? Well, one of the things that happened was that I posted on social saying that, you know, Musk should stop and stop Grock from being able to do this. And then, of course, I became a victim of the same trend.

So at Grock put this woman in a bikini and most of the images were sort of pornified, I would say. They weren't nude images, but they were me with inflated boobs in a bikini. Then people tried to make it funny, I guess, by putting a swastika on me, all hammer and sickles, depending on my friends, or making me pregnant with quintuplets and putting me in a kitchen, because, of course, that's the only use for women is birthing and making dinner. But the really disturbing one was a video of me being surrounded by a group of men, chloroformed over my mat, and then I fall backwards onto a chair and then my skirts are lifted up. So no nudity, but clearly the next frame is meant to be a sexual assault. So it was really, really distressing and disturbing.

that was the point at which a friend said you probably need to reach out to a lawyer because each of each of these things we reported to x none of them contravened their terms and conditions and we just keep getting those sort of repeat emails coming back we've looked into this and there's there's nothing to see here so I did engage a lawyer who sent a letter a letter to x to ask for these to be removed and they were but it was at that point that he said actually I think you might actually have a case we think that xai have broken i mean i think it's brilliant that you did that and i can see the through line of your legal training and um how practically you tackle things but when you were just telling me that it really turned my stomach i had a very physical reaction to you describing what they did and i i i just i wanted to sort of find out more about how it felt I mean I'm really impressed by how how you how you handled it and and and the

sort of default to action but it must have been really awful it it was it was awful it but this is done to demean to degrade and to push women out of online spaces from my perspective I tend to use bad or difficult things that happen as a way of thinking about how do we change this for the future my old boss was an amazing woman called Tessa Jow and she used to say to me if anything went wrong she would say how do we change this negative energy into something positive how do we how do we work our way out this into into a place where we actually feel that we've learned from it and so I guess she instilled in me that sort of sense that when something bad happens you've got a choice you can you can you can wallow it in it you can try and forget about it or you can use it as a catalyst for change and so I was quite trepidatious about the idea of

taking on a legal action not least because of the person who is at the top of xai now space xai he's the world's richest man but also because I didn't want myself to necessarily be the story many many women had things that were created of them through both grok but also other notification tools that are much much worse having said that i knew that if there was the opportunity to say that actually they've broken the law there was an important point of principle on behalf of all of the women and the children whose images were manipulated and misused by grok who would never ever been able to bring a legal case and so I decided that I wanted to do this because I wanted to be able to show that no one is bigger than the law. It doesn't matter how rich you are or how powerful you are or how big a reach you have in terms of your social media presence or

who your friends might be at the top of various governments. If you break the law in our country you need to feel the force of that and as as an ex-lawyer and now as a legislator it is important that we uphold our law and so it's also about giving control back I think to us as individuals but also as a society so many of these digital changes have been done without our consent our say and I think there's a feeling that there's a sort of runaway train that we're all on whether we want to be or not because we all need to use online spaces for our work for connection and so my view is that we should be expecting companies to respect our law or at least make new laws to bring them more into line with what society really needs. You said you gave it some thought in that moment of trying to decide how to respond what are you weighing up who are you talking to are you going back and talking to your team or your family to work out how you how you should respond

knowing that it may invite further speculation and further attention my family mostly so my husband in particular it's one thing being a politician and arguing your case to change the law on behalf of women and children it's another having your own legal case and so I was worried about the impact that that might have on my family, my children. I'll be honest, because I have been, you know, at the heart of pylons on X before. I also know that this can bring danger to politicians. And so I worried about the fact that I am obviously engaging and making the defendant somebody who has got this very wide reach on social media. However, I don't like bullies. and I very strongly believe that it's important to fight on the field that you've got and this was an opportunity I felt to make a difference.

So I weighed that up and thought that it was worth having that kind of attention spotlight and potential danger because there is a bigger issue here that we need to get right. When you say you don't like bullies, are you talking about the, let's face it, men who used Grok in this way in the first place, or are you talking about Elon Musk? oh i'm talking yeah i'm talking about yeah the men who use he use this as a form of trying to push women out of the online space and and actually it was quite clear and quite lost the comments that they were saying well if you don't like this then get off the platform either that or if this is too hot for you get out the kitchen which i always find quite funny because they put me in the kitchen it's but like do you want me in the kitchen or outfit and i think it's really interesting when you look at X in terms of its demographics in the UK it's highly weighted towards men so women are already voting with their feet in relation to this particular platform

and I think some of that's down to the fact that there is just so much misogyny and abuse on it. So I'm assuming correct me if I'm wrong that the other side of the response to going public and deciding to continue with being public on this issue and to fight it in court is that other women have come forward to you did you did you have other women coming forward and talking to you about your decision to to make this legal absolutely and actually that has been one of the I guess the heartening things has both been people coming forward and saying thank you because this has now given me the license to actually accept that what happened to me wasn't my fault it's not my shame and so in in a way what I'm hoping is that this has helped some victims to process what has happened to them and understand that this was not right it's not right I've also had emails from parents where they've said that this has happened to their daughter it's been done by boys at school I've had teachers saying that boys have made pornographic

AI imagery of them and it it is actually quite rife and it's something where we've just sort of let it happen and wash over us and accept that somehow this is now a part of the society we live in so yes quite a number of people have come forward obviously not not all of their experiences have been through grok and i think one of the one of the things is is that often when you see these films or images you don't know what platform they've been created on i mean the only reason why i knew it was grok was because it was integrated with x and you could and you could see the commands that have been given by the users. But much of the time I think when this content is created most victims don know by who or on what platform and so there really very little that they can do about it and I think it that sort of sense of helplessness that you have been violated and that is the word that comes up so frequently

from victims of this this particular issue and it's not just that it's nudified of course much my content wasn't it's the lack of consent and lots of people say oh but you know it's just it's it's fake right it's just a laugh but what people don't understand is that it is hyper realistic and it's your face it's it is peace of you peace of your soul it's not just pixels on the internet it is a person and because obviously I've worked for a long time in violence against women and girls and spoken to many victims of sexual assault and to children who've been victims of child sexual abuse it is the lack of consent that is so disturbing and it can leave people with PTSD and quite deep trauma and there have been a number of people I've met who have said that the world seems less safe to them because this has happened to them.

I mean, I think you've already touched on it, but what is the biggest myth around deepfakes that you would wish to dispel, that you now, through experience and through the people that you've spoken to, wish to dispel? That it comes without any harm. That it's just a laugh. That it's just a laugh. There is clear harm. And even when it's not done to demean or degrade, even if it's done sometimes by friends. The fact is that you haven't consented to that happen. And there is something about the human psyche, which is that when you see yourself so realistically created in a different situation or scenario, something where that's not something you would have done, I think it's quite discombobulating. I think your brain can't quite process it because your brain knows it's not true, but it looks very true. And I think psychologically, we still don't really understand how this content is going to change the way that we think about our world, how we know what is true and what isn't true.

Now, of course, you can do fantastic things with generative AI. And all I'm calling for is that we respect people's consent. And that's why one of the things I've been working on since this happened is to look at whether we could create a system of consent to be manipulated by AI. Could we create a greater right over our likeness online so that there's a new social norm, which is that, no, you can't copy and paste that picture into an AI chatbot and do whatever you want with it. That's not yours. It's not your data. it's their data and you're breaching the law if you do that and I think were we to do that would have a wide range of applications it wouldn't just be in this space of the sexualization of women and children but also with political disinformation and we are we are so close to a moment where somebody produces a hyper-realistic video of a politician saying something that they

didn't and then that leading to real world harm and so we've got we've got to find a new social norm that says unless for public figures it's parody or satire absolutely got to protect that but in all other instances your picture your voice and your moving image is your personal data and it's not for anyone else to use so that's interesting because you took you touching there on new social norms and the possibility of new laws to keep up with this technology. But your lawsuit is specifically under a data law. Talk to me about that versus a law around consensual imagery or deepfakes particularly. Yeah, so my case is brought under existing law, the GDPR, the data protection law, and it contends that Grok misused my data without my consent. It didn't have consent to process my data, the data being my face, my likeness.

But also under privacy laws that I had a reasonable expectation of privacy and that particularly the video of me being prepared for sexual assault reaches a criminal threshold and that nobody should be expected to be on the receiving end of a product. Let's remember, this is actually a product that has been designed by software engineers and developers, and that actually products should not lead to an invasion of privacy and the harm that I then experienced as a result of that, including distress and a sense that I was being harassed by this particular creation of this particular video. So it rests on good, solid legal ground, and we await the defence from SpaceX AI. So that was going to be my next question. I don't know how much you can say legally. I'm not sure how you're bound, but what process, what part of the process are you in?

What happens next? What's the timeline for a resolution to this case? so like with a lot of legal cases it's very very hard to pin down a timeline but my understanding is is that we we should be receiving the response and the defense from spacex ai in in the near future but we don't have a deadline or a particular time that i can put on it after that my understanding is is that the case will then be looked at and if there is a case to answer then it will go to a court hearing but once again no idea when that might be and I think there's lots of legal things that I don't understand that need to be worked out in the meantime as well between between my lawyers and their lawyers. And have you ever had any acknowledgement on X itself from Elon Musk or anyone at X about this case? Has he ever spoken publicly about it? No. Would you expect him to? I don't think so. They clearly want to challenge this case.

And so the best place to do that, I think, is in the courts and to leave it to see whether the merits of the case come through. Take your curiosity with you on KUOW's new podcast, Pocket Science. In each episode, we catch up with local scientists in the lab and in the field to find out how they're pushing the boundaries of what we know. From the surface of an icy moon of Saturn to the inside of a human brain on psychedelic drugs, The answers might surprise you. Listen to Pocket Science on the KUOW app or wherever you get podcasts. AI systems don't operate in a vacuum. They need context. And that context comes from data. Choose the wrong data and your AI could deliver something between hallucination and guesswork. Dun & Bradstreet Context Layer, anchored in the Duns number, delivers the context your AI systems need

to reason, decide, and act across your workflows. Business identities are resolved. relationships are clear and continuously updated data keeps those systems grounded so your ai doesn't guess and you can operate with confidence give your ai the context it's missing learn more at dnb.com trust your ai ai can make businesses faster it can also make mistakes expose data and execute the wrong instructions faster threat locker helps organizations set boundaries ai cannot ignore by controlling what applications can run what they can reach and how they interact with sensitive resources. Innovation needs more than good intentions. It needs limits. Control AI with zero trust at threatlocker.com slash wired. That's T-H-R-E-A-T-L-O-C-K-E-R.com slash wired. You were in New York to address City Council about your case. You didn't get to speak to anyone

directly at SpaceX, is that right? No, I mean, despite being subpoenaed, they didn't turn up. So four of the other frontier AI companies did attend under subpoena. SpaceX AI didn't, but they eventually did get their hearing with SpaceX in the city council. So they have now had their opportunity to put questions and indeed did raise my case with them, I understand. Is it a little enraging to attend a hearing like that, a company that you're suing who don't even think it's worth, who don't bother turning up? I think it's quite rude to the elected members of New York City Council that you are a business who operates in New York. And indeed, I was told that they were just four subway stops from City Hall. so I had travelled quite a long way from England to be there and they couldn't be bothered to turn up which, yeah, it felt arrogant and a bit of a snub

but they've now got there so that's a good thing. Since you made the decision to take this to the courts there has been a number of proposals in Parliament notably that there was a proposal for a kill switch bill at one point though I believe it dead or it not going very well Is that right So, yes, there have been proposals put forward for a kill switch and my understanding is that the government rejected that. There continues to be strong calls for regulation of particularly superintelligent AI. we have we've heard various theories whether it's 2 percent 10 percent 20 percent risk of human extinction from this this sort of race to super intelligence a recursive AI and we wouldn't allow those margins of risk in any any other sphere we would not allow it in bridge building or or airplane making, or they didn't even allow it when they were testing the nuclear bomb and the Trinity testings.

It feels to me as though so much of this decision-taking is outside of democratic structures. It's not with the public. It is being done by some incredibly, exceptionally wealthy companies and organisations who say that they have the interests of the planet at heart, But each time actually they're asked about whether that is the future of humanity or whether they think that these risks could kill humanity, there's a sort of shrugging of the shoulders and we'll just keep going because we want to make money, I guess. Well, part of the problem, it has been reported in the British press that there are members of parliament, some of whom who are working on committees to do with technology, who still don't really understand what AI is. And clearly you're not one of them because you're a sort of early adopter of the dangers of this issue. But do you feel, working across parties, that you're still having to explain technology and its dangers and possibilities to people who are making laws or should be making laws to protect the public from AI?

Well, look, I don't think that's only the case about AI. We are humans after all, and we have to deal with every single subject that exists under the sun. We have to get our heads around things that we've never heard of before. A bit like being a journalist, actually. Yeah, that's true. But you don't have to then make laws and decisions. So, you know, I don't want to be unfair to my colleagues. You may be a specialist in housing policy, and then you need to get your head around AI. But I would say that this is the most important issue in front of anybody in parliamentary politics at the moment. And therefore, everyone needs to get their head around AI. I'm not a technology expert. My background is in social policy, women and children. But because we did quite a lot of work on the Online Safety Act when I was working for a children's charity, we campaigned to get the age verification ban for children in pornography into the USA. and indeed created quite a strong piece of legislation in respect of that we had to get

our heads around online safety on the online space the problem was of course was that act didn't include AI chatbots which there were a number of members that tried to get it in and we would have been a much better place had we actually succeeded but you know I'm still not a tech expert even though I had quite a while to try to understand these things but my view is is that this is definitely the new front line of where harms are occurring we can see that hugely an increase in ai generated child sexual abuse material which has quadrupled right it is absolutely exploding and of course it all comes down in the end to the people who create the products and that's why we should be seeing digital product regulation in the same way as we would any other product regulation My colleague Tom Collins has pulled together a brilliant white paper on digital product regulation. I'm very, very hopeful that the government are going to look at it really, really closely because there are always design choices that you can make and that can help to then prevent these harms.

These harms that are completely knowable. We can predict these harms and so it should be easy to design them out from the start. That is a common refrain and that was at the heart of Starmer's message before he left office, compelling companies to do this themselves. They can do it. Why don't they? Profit. But also, let's be honest, these are global tools and companies. And so unless you have international agreement, it is a bit like playing whack-a-mole, isn't it? I think we've just seen that Ofcom have had to sort of drop their case on 4chan because essentially 4chan appealed in America. And it's one of those things where if you can't get the international architecture right, you can have in-country laws and you can enforce them, of course. But that's always up until a point where actually the person who made that horrible video of me being chloroformed lives in the US.

So the police aren't going to follow it up because they're out of country. Do you know exactly who it is? You know who did it? I don't know who they are, but the police did find who they were because it's a criminal offence. What they did, they just weren't able to bring the charge. We've spoken in the last 40 minutes about really the full spectrum of human awfulness. I don't think there's any area of it that we haven't covered. Sorry, welcome to my life. Well, I do all the dark things as I tell people. I'm really struck by how cheerful you are despite this. How? How do you remain so upbeat, hopeful, dare I suggest, while dealing with all of this information? Because you've always got to have hope for the future. And I do have a really strong conviction that we, the right group of people pushing towards a goal, you can achieve that. And you don't have to accept that things will always be like this.

that when people come together and when communities come together you can act for change and I've I guess I've proved in some of the things that I've been able to introduce into legislation the we're just about to introduce our ban on strangulation and pornography which is another thing that I've been campaigning for and introduced an amendment last year which the government accepted there are always things that we can do to change you just need to keep going and also I would say that when when you work with these terribly dark areas yes it can it can be quite vicariously traumatizing but actually the people who work within the sectors are some of the most hopeful people partly because if you lose hope then then then it is it's really bad and we can't because we're talking about protecting vulnerable children and we are fighting for women and our right to be free of this harm.

Thank you, Jess. It was really great to speak to you today. Thank you so much for your time. It's really lovely to meet you. The Big Interview is produced by Lina Macitsis. Our executive producer is Kate Osborne. Music and mixing by Pran Bandy. This episode was fact-checked by Matt Giles. And I am, of course, your host, Katie Drummond, Wired's Global Editorial Director. Check back here on Thursday for the latest episode of Uncanny Valley, where Wired's writers and editors add you to their Slack channel. Take your curiosity with you on KUOW's new podcast, Pocket Science. In each episode, we catch up with local scientists in the lab and in the field

to find out how they're pushing the boundaries of what we know. From the surface of an icy moon of Saturn to the inside of a human brain on psychedelic drugs. The answers might surprise you. Listen to Pocket Science on the KUOW app or wherever you get podcasts. Are your AI initiatives underperforming or just under-informed? When you depend on agents to reason and act, you need trusted, business-specific context grounded in reality. The Dun & Bradstreet context layer bridges the gap between your existing stack and your AI, delivering verified data across your workflows. Business identities are resolved, relationships are clear, and continuously updated data keeps agents grounded. So your AI doesn't have to guess, it reasons. And drives decisions you can rely on. Give your AI the context it's missing. Visit dnb.com slash trust your AI. AI can make businesses faster. It can also make mistakes, expose data, and execute the wrong instructions faster. ThreatLocker helps organizations set boundaries AI cannot ignore by controlling what applications can run,

what they can reach, and how they interact with sensitive resources. Innovation needs more than good intentions. It needs limits. Control AI with zero trust at ThreatLocker.com slash wired. That's T-H-R-E-A-T-L-O-C-K-E-R dot com slash wired.