Committee on Women’s Rights and Gender Equality Ordinary meeting HEARING Hearing on AI, gender-based violence including sexual images of women and children: the case of Grok
Legend
Transcript markers
- text Sentence being played
- text Search result
- text Intervention cited by the link you opened
Click a sentence to play the video from that moment. Select text to quote it with the speaker, timestamp and link.
More options
FEMM European Parliament
Follow its sittingsSuvi Uski
Thank you, Chair, and good afternoon, Parliament. My name is Suvejewski, and I'm the CEO and co-founder of Somebody. We protect hundreds of thousands of school children and teachers and employees in the Nordic countries, Finland and Sweden at the moment. Before somebody, I have worked as a social media researcher, as a doctor of social psychology, trying to understand the developmental use of social media platforms. So I have a wide experience in both, like, seeing the social media platforms evolving before COVID and after COVID, but then also seeing how the online crime and abuse of children and adults are happening in the different online environments.
Today, I'm going to share a little bit of our expertise with you, and I stay at your disposal for your questions later. Let's move to the next slide, please. At Somebody, we provide something unique to our users. So, as said, we protect hundreds of thousand school children. And that means that we offer the best possible expert help.
In my team, we operate an online safety emergency unit that has professionals from law, psychology, social psychology, and also some cyber experts. This means that we can actually give the best tools for our victims who are seeking help from us, in order to calm things down or trying to minimize the risks or consequences that are happening to the victims. Also, noteworthy to mention that we are a trusted flacker by the Digital Services Act and also that we are a commercial organization, which means that the schools and municipalities, campuses, and workplaces can acquire our services, and the end users, the members of those populations, can use the service free of charge, which makes the use of service a very low threshold approach to the online victimhood. Let's move to the next slide, please. I brought you here today Jenna, who is a 15-year-old victim of sex torsion.
That's a very typical case from our daily operations at our emergency unit. One might here think, okay, how is this sex torsion happening? But unfortunately, it's very straightforward. What happens to Jenna is that Jenna receives a Snapchat direct message, DM. And the message says, as the small print goes on the slide, you pay or we publish an AI porn video of you.
As we all here know, when we talk about deep fakes or all kinds of AI generated material that are harmful or illegal. We also know that it's very difficult to distinguish between real and fake materials. And for Jenna, if we think of the psychological or other consequences, they are real, even though the material would be fake. So, and if we look at the users of our online safety emergency unit and label the psychological consequences of being a victim or a target of this type of sextortion act, what does it say? It says anxiety, depression, self-harm, and suicidal behavior.
And for Jenna, it means, in practice, a very shameful, humiliating feeling and the fact that it's very hard to do anything. It's very hard to go to school. It's very hard to partake in any activities. What is concerning here is that the situation is unbearable, and we don't have many good means to help Jenna out here. And what is also something that we've witnessed at somebody is that even though deepfakes, in many cases, are made or produced for entertainment, what we see is also for a crime, for sextortion, to make money out of this type of illegal business.
Let's go to the next slide, please. I brought you also here an age histogram of the cases from Finland from the last year that we have tackled with the school children, from 6 year olds to 17 year olds. And you can see that we serve kids of all ages when it comes to school children, so we actually know what is happening out there. And I am very much available for sharing our views, and these are available publicly on our websites as well. Let's go to the next slide, please.
And to wrap around a little bit, I also wanted to bring you some screenshots from our Swedish operations. These are in Swedish, but easily translatable. It says, can you make me a nude picture of a schoolmate, of a friend, and these are like sent, these are messages sent to us in a trolling manner, but also at the same time depict the culture that is evolving around this scene of deepfakes. So it's just for fun to fabricate these pictures on a daily basis. What comes with this is also that we talk about crime that is evolving around the school children, is that when we think about victims of deepfake, the victims are not only the victims who are targets to these deepfakes.
Victims are also these 10-year-olds who are producing this illegal or harmful material on a daily basis or consuming it. It's not healthy for one's psychological development or sexual development to be in interaction with these types of materials in the long term. Okay. Then, to the last slide, please. And then also, if we look at the cases from our emergency unit, the numbers when we ask, like, have you told a trusted adult about this situation before?
Are pretty high for the cases of sextortion and extortion. So, we know that in 71 percent of the cases, we have been the first stakeholder to hear about the situation of that young person. And I have to say that in other types of online crime cases, the percentage is somewhere between 30 to 40 percent where we are the first stakeholder. So, this also underlies the fact that when there are victims, and as UNICEF in Japal Eckpad's study just revealed, it's like 1,200,000 children who have already been victimized in 11 countries. You know, what we're seeing is that these numbers are just growing.
Thank you.
Lina Gálvez
Thank you. Thank you very much. Indeed, impressive, the way you are doing and also the numbers and the reality, that it is going around. And now, online, with us, I would like to give the floor to missus Sei Kiwowo, who is joining us remotely, as global adviser on digital governance, safety, and institutional accountability. She's also the founder and former CEO of Glitch.
And she works at the intersection of digital and artificial intelligence system safety, public policy, and institutional design. Her focus is on advancing a public health approach to emerging technology systems, operationalizing duty of care structurally before harm happens, and not after. So, are you online already? Hello?
Hi. Can you hear me?
Speaker 4
Thank you again for having me. It's been good to see you. It's unfortunate that we're having this kind of conversation.
Lina Gálvez
Yes. Indeed. Indeed. But it is necessary to have it as I'm sure you agree with us. So you have the floor for 7 minutes.
Thank you very much for joining us and accepting our invitation. So the floor is yours.
Segi Akiwawa
Thank you. Thank you so much for having me. And, you know, I was at the committee in person in Brussels at the 2024. We were talking about the kind of next steps post regulation, what we needed to put in place. And it's a shame that, well, 18 months after that conversation, we're still seeing tech companies fall behind on the basic standards of care.
So this is why I'm really delighted to be joining you remotely here from the UK. Chair, members of parliament, thank you for the invitation. I want to start with a simple truth. What we are seeing is not unexpected harm. It is unstructured harm.
When harm is unstructured, institutions cannot respond consistently, and accountability becomes difficult. Europe has already taken global leadership in this space through the Digital Services Act, the AI Act, and GDPR. You have established an important principle that digital systems must be governed in the public interest. But we are now entering a second phase, a phase where it is faster, is more scalable, and is increasingly designed into systems, not just emerging from them. The case we are discussing today, AI-generated sexual images including of women and minors, is not an isolated failure.
For years, the nudification of women and girls has been a known and documented problem online. In 2023, it was reported that 34 Nudify app providers attracted more than 24,000,000 users. These tools were trained almost exclusively on the bodies of women, and in many cases, didn't work on men at all. We also know that around 99% of people targeted by sexual deepfakes are women. So the demand was known.
The risk was documented. The precedent was clear, choosing to launch a mainstream AI product and one that could nudeify women and children in plain sight on X without any guardrails to prevent this use case was not an oversight. It was a design decision, likely one motivated by pure profit. It shows us three structural gaps. Number one, platforms can generate harm at scale before safeguards are in place.
Two, there are no consistent ways of classifying severity across systems. And three, victims are still expected to navigate fragmented reporting pathways. So this creates a situation where harm is so visible, but responsibility remains scattered. In preparation for this discussion, I analyzed a small but highly diagnostic evidence base from movement leaders and politicians globally who've undertaken my training and capacity-building workshops. 22 responses from women in public life from different continents around the world and democratic structures, as well as leaders in the refugee and migrant sectors.
These two communities that we know are the most vulnerable and disproportionately impacted by AI and tech-facilitated violence. What this sample shows us is that simply the doesn't just exist. It shows us that visibility creates foreseeable risk, but institutions do not activate a consistent duty of care in response. Harm is not random. It's clustered around leadership, public advocacy, elections, funding decisions, and moments where individuals dare to challenge power.
Across both datasets, the pattern is consistent. Institutions notice harm way too late, treat it as a reputational interpersonal issue, and fail to classify it as a governance risk in public life, where most leaders reported either no response or delayed reaction. In migration refugee context specifically, harm was strongly racialized, yet institutional protection depended on individual maturity, not on a shared standard. So the conclusion is clear. Digital harms become institutional failure the moment the risk was foreseeable and no duty of care was activated.
So we need to move the conversation today away from content moderation to duty of care, away from reactive response to preventative systems, and from individual reporting to institutional accountability. Digital harm now meets the threshold, in my opinion, of requiring a public health approach in how we now regulate and ban exploitative tech. In public health, we don't wait for harm to escalate before acting. If something is predictable, repeated, and a pattern, we treat it as a system failure, not an individual issue. Therefore, what is currently missing and what I invite our conversations to center around today is not talking about awareness, but about the infrastructure.
For the ability for institutions to do three main things, which are on this, on my slides today. One, identify harm early, classify it consistently, and respond in a way that is proportionate and accountable. This is the work I'm now developing through what I describe as building digital power infrastructure, not as a product, but as a systems approach, meaning that we have a four-tier model, a shared way cross-country to document harm safely without retraumatizing individuals, clear classification on the severity and escalation so all institutions know how to act. Let's raise the standard and the bar across all sectors. Governance protocols defining responsibilities and expectations of platforms, and most importantly, pathways when people become a victim of tech.
This approach directly builds on the direction that Europe has already set. So the question now is how to take this four-tier model so that harm is measurable and actionable in real time. So to close, the question is no longer should platforms be responsible. The question is, do our systems make responsibility possible? If we want safer AI systems, we need systems that translate harm into institutional actions.
And I would be very much happy be glad to continue this conversation with you. So thank you so much for the invitation and your time this afternoon.
Lina Gálvez
So, thank you. Thank you very much, especially for pointing out the systemic nature of the problem and also why we have to give a systemic answer. We have done a lot already, but definitely, it is not enough precisely because it is systemic. And now, I would like to give the floor to mister David Rangel, head of data and digital sector within Justice Digital Immigration Unit at the fundamental rights agencies. So thank you very much for your help, and the floor is yours for 7 minutes.
David Rangel
Thank you, chair, for the floor and for inviting the EU Fundamental Rights Agency to this hearing. I'm tackling a very important topic touching on several fundamental rights. In my presentation, I would like to highlight several findings from reports of the agency that speak to the topic of today, starting with the scale of gender based violence in the EU, fundamental rights risks when developing and using AI, gender based violence online and online content moderation, and lastly, considerations related to deepfakes when they are used for violence. As we all know, gender-based violence is a structural problem. The agency published a report together with the European Institute for Gender Equality survey data that were collected by Eurostat and the fundamental rights agency in eight member states, plus Eurostat covered the entire EU data on experience of gender-based violence.
Unfortunately, the results confirm the scale of the problem that the survey highlighted already ten years ago. So almost one in three women experienced violence since the age of 15. Nineteen percent, one in five, experienced it by a domestic partner. 17 percent experienced sexual violence during adulthood. 6.6 percent reported to have experienced sexual violence in childhood.
Speaker 7
Thirty point eight percent, so almost one in three, also experienced sexual harassment at work, where seven percent reported that it was done with the use of online tools, and 9.2 percent have been raped. In this survey, the European Institute for Gender Equality and FRA added questions in those eight member states that we covered to add also more information on gender-based cyber violence. Here, we found that almost 1 in 2 women experienced sexual harassment outside work. Almost 10 percent, 12 percent experienced cyber harassment that includes either direct messages to them or being approached on social media. 8.5 percent experienced cyberstalking, which mostly happens online.
And lastly, 1.1 percent of women interviewed in those eight member states indicated to have experienced abusive behavior involving intimate images or videos of women interviewed, either real or manipulated in a way that caused harm. Now 1.1 may not sound like a large number, but overall, if you apply this to the entire EU, there are 230,000,000 women in the EU, and 1.1 percent translates into two and a half million women. So that's the scale of violence against women. Now we have AI, more and more, AI used in different areas. But we also already know that technology is not neutral.
The Fundamental Rights Agency's report of 2020 already highlighted whenever AI is used for whatever purposes, you have to be very careful and see if it interferes with privacy and or data protection, if it goes against equality or nondiscrimination principles. It may affect access to effective remedies if opaque systems are used for decision making, and virtually any fundamental right may be impacted depending on the area of use of new technologies. As all of you know, there's very active policy-making to address such risks horizontally. One example when AI is used that has also a strong gender component is biased algorithms. At the agency, we published a report in '22 that looked into how speech detection algorithms may be biased towards ethnic minorities on the one hand, but also against women on the other hand.
What we've seen when we tested it, that was only based on text, but we also tested it in English, Italian, and German. And in Italian, you can also see how gendered nouns impact the algorithms. So we did not develop these algorithms to use them in practice, but to test them with invented words and sentences, and see how often they overreact to non-offensive sentences. And the example you can see on the slide, for example, is that if you use the feminine version of Muslim in Italian, it overreacts much more strongly than when you use the masculine version. This is just one additional dimension of when you use AI that has a gender bias, which translates into problems when you use it for detecting certain hate speech.
David Rangel
Gender-based violence is a particular problem online, and it's difficult to moderate it in practice. We published a report in '23 where we collect the data on social media and try to find how easy it is to find hate speech that slipped through content moderation systems. Again, that was text based, but still would like to use the results to show the case that compared to hate speech targeted at ethnic minorities, it was the vast majority of posts we found online was targeted at women just using simple keywords. And a lot of the language we found used denigrating language, offensive language, but also a lot of negative stereotyping. In practice, and that also translates to the topic of today, there are challenges encoding this kind of content.
Speaker 7
So coders were sometimes uncertain for certain texts. Is that hate speech or not? And more importantly, is that illegal or not? And this translates into the challenges of moderating content online nowadays. As you also know, there is no silver bullet.
There's no single solution to this. And the Digital Services Act, we heard about it, has several responsibilities, including targeting public authorities, individual users, opportunities to report content. Civil society organizations play an important role, such as trusted flaggers. But, of course, platforms also have to take proactive measures and need to assess the risks of their systems. Before coming to the topic of deepfakes, I would like to highlight fresh data that the agency published just 4 weeks ago.
David Rangel
So we asked people in the EU in a EU-wide survey about their concerns related to child safety online, and that's a striking result that almost everyone, between 80 and almost 90% of adults interviewed, are either very or fairly concerned that children, when they use social media, encounter violent content, may be approached by people who want to exploit them, may spend too much time on certain apps and become addicted, but also that they may be exploited based on personal information they share, where every second person is very concerned about. That they may encounter pornographic content. Also, terrorist content may be used to commit a crime or encounter conspiracy theories. So as a last point on deepfakes, as one of the areas that may be used to exert gender-based violence. We heard it already before, but I thought it's good to acknowledge how deepfakes regulated in the EU, especially when they're used for gender based violence, is fragmented at the moment.
But that is also good because we have different laws that tackle the problem from different angles. We heard about the AI act that has transparency requirements included when AI is used. However, so far, it seems at the moment that it does not fall under the prohibitions. The Digital Services Act, of course, is a very important law that tackles the responsibility of very large online platforms in their systematic risk assessment. This topic has to be tackled for sure.
We also already heard whenever personal data are involved, the GDPR applies in full as well. Lastly, we have the violence against women directive, which criminalizes nonconsensual sexually explicit deepfakes. The transposition is due only in June year, but some member states already criminalized deepfakes and went ahead there. Of course, there's also the directive on child sexual abuse material that criminalizes this material online. What is, of course, important to clarify in the discussions you have is what are the liabilities of providers, those putting the tools on the market, and, of course, what use of this tool is criminalized and illegal.
Speaker 7
I went a bit overtime, and just wanna say the agency will continue to provide more data and analysis on online content moderation and AI, more specifically also on this topic. We suggest the strong implementation of existing frameworks relevant for the protection, and we need increased transparency and knowledge needed for the enforcement of existing laws. Thank you.
Lina Gálvez
Thank you. Thank you very much. It was important to listen to all of you if we took a bit more time. And now we have a video from our colleague, Brando and Ife, that was the rapporteur for the artificial intelligence act, is the Inco co-chair and the Inco Liber Working Group of Artificial Intelligence, and he has recorded a video for us. So go ahead.
Thank you.
Brando Benifei
Thank you, Chair Galves Munoz, and thank you to the FEM committee for organizing this hearing. I regret I cannot be there in person, and I'm grateful to the expert speakers for the work they do and for the testimony they bring today. What makes AI so dangerous in the context of gender-based violence is not just the nature of the harm, but its scale. Deep fake videos targeting women have increased exponentially, with 99 percent of individuals targeted being women and girls. AI industrializes misogyny, removing the barriers that previously limited who could produce this content and how much of it could exist.
The Grock scandal illustrates this with particular clarity. Over half the images generated in just one week depicted individuals in minimal attire, 81% women with a majority of apparent minors included, produced through simple prompts with no circumvention required. The central regulatory challenge here is genuinely a difficult one. AI systems simultaneously generate and moderate their outputs, a paradigm shift that frameworks designed around intermediary liability have not fully absorbed. The DSA creates real obligations for X, including the explicit duty to conduct an adult risk assessment before deploying features with foreseeable risks of gender-based violence.
When X rolled out Grox image capabilities without an assessment, it violated the existing law. The AI act also requires frontier models to mitigate these risks, but still leaves too much to provide us at discretion, which is precisely why parliament work in the digital omnibus to add the prohibition of new defecation systems alongside clearer expectations on what effective safeguards must demonstrate. The prohibition matters, but what is clear to me, though, is that the priority right now is to enforce the rules we already have. The commission investigation into acts must produce consequences that actually change design decisions, not ones that get absorbed as a cost of doing business. The millions of women harmed by systems like Grok cannot be treated as an acceptable externality.
We have the rules. We need the capacity, the resources, and above all, the political will to make them real. I wish you a very productive discussion. Thank you very much, and see you too.
Lina Gálvez
Gratuit, Brando. Thank you, Brando, even if you're not listening to me. And now, I would like to welcome the co rapporteur from Inco and Libbe. Now, also a member from our committee, Arva Cocollari. She's the IncoCo rapporteur on the artificial intelligence omnibus.
So thank you very much. And the floor is yours for two and a half minutes.
Arba Kokalari
Yes. Thank you very much, chair. Hello, everybody. Thank you to our speakers, and thank you for this initiative, a great initiative for this hearing today about this growing problem that we have on gender based violence online and now in the era of AI. Online abuse can harm women's dignity very, very quickly.
We see that every day. And it may begin with an ordinary image, such as a vacation photo, which is then manipulated to undress or sexualize her without consent. And this is a problem that affects many women across Europe. And women in public life, for example, including elected representatives, are often particularly exposed. We have such an example in my home country in Sweden, with our vice prime minister who has been exposed like this.
The chatbot Grok on X has reportedly allowed users to create AI-generated sexualized images of women without their consent. And according to the reports, Grok has created and published 4,400,000 images in just nine days. Because of this terrible violence that many women have experienced, I am glad and it's good that the commission is now investigating how these platforms can be held responsible. Because it's very important to say that we actually have laws in Europe on this. We have the Digital Services Act, DSA, the GDPR, the AI Act, the violence against women directive, the Istanbul Convention.
It was mentioned here. I myself, I've been the rapporteur or am the rapporteur for the Istanbul Convention. I was a shadow rapporteur writing the Digital Services Act. And now I'm the rapporteur for the AI Omnibus. The laws that we have in Europe need to be enforced.
So we are now in the negotiations in the AI Omnibus, also looking into how the AI omnibus can help to a safer environment online, where women are protected from this kind of violence. Because it's also important to say that the solution is not a ban on technology, which some voices seem to promote. That would be as inefficient as banning cameras just because people use them to take nude pictures without people's consent. We are living in a digital society. But it's important that the laws that we have in Europe are adequate in where and what they can be used for, that they are properly enforced and trustworthy.
This is something that we are currently discussing in the AI omnibus. And I have a question for our speakers. How can the EU become better in using the existing legislation effectively so that we can stop these abusive deepfakes without banning the technology itself.
Lina Gálvez
And also for us starting the questions to our experts today. And I would like to give the floor to the other co-rapporteur on the artificial intelligence omnibus, Michael McNamara, who is also the Liebe co-chair of the IncoLiebe Working Group on artificial intelligence. So thank you for coming. Thank you for being here, and the floor is due for two and a half minutes. Thanks.
Michael McNamara
Thank you very much, chair. Across political groups, across institutions, legislators looked at the scale of nonconsensual sexual imagery as what tools like Grock's Spicy Mode, but it's important to say this. It's not just about Grock's Spicy Mode. There are many other modifiers out there. We looked at what they were doing to women and girls right now, and decided that this was a moment to act.
I think that must be acknowledged. It's a testament to both the council and the parliament and groups across the political spectrum of this parliament that, when the opportunity arose, it was taken. For many groups in the parliament, including my own, securing a ban on unification and undressing applications is a key negotiated outcome that is sought from the AI omnibus. We fought for it because nonconsensual intimate imagery is a systemic, gendered harm being industrialized by AI. But I think we have to be honest about where negotiations currently stand.
I am worried, not about whether we will get a ban, we will, but I'm worried about whether the ban will actually do what we need it to do. Under one version of the text currently on the table, it states the prohibition should be limited to realistic depictions of exposed intimate parts. In particular, the genitals, pubic area, anus, fully exposed buttocks, or fully exposed female breasts, including nipples and areolae. So with this wording, placing a star or a bull's eye over some a young person's or even an old person, any person's genitals would be sufficient to take the image outside of the scope of the prohibition. What would be required for the prohibition to kick in would be full nudity.
I don't believe that is the ban people are seeking. I don't believe it's a ban that will actually be effective. Insofar as it is a ban, it is one with an open instruction on how to circumvent it for those companies that wish to monetize this activity. And, clearly, there are economic actors, albeit a small number and on the fringes, who do wish to monetize this. We can also, of course, look at what other jurisdictions are doing.
Only last week, in the United Kingdom, the UK's ban was promulgated, which went much further and referenced all or parts of a person's exposed genitals, buttocks, or breasts. That is important because it prevents the circumvention of it by just, you know, putting in a token, as I say, bull's eye symbol, star, etc. So we now have the opportunity in this trial to set a standard that actually protects women and children. I would urge colleagues to hold the line on a broad, anatomically precise definition, not because we want to be difficult, not because we want to stop progress, not because we want to stop economic activity, but because the victims of this behavior deserve nothing less. And we need to reduce the number of victims in the future, not provide an instruction to tech companies on how to carry on regardless.
Thank you.
Lina Gálvez
Thank you. Thank you very much. And then, and now, we will start with speakers from different political groups. Starting by EPP, Eleonora Meleti has the floor for two and a half minutes.
Thank.
Eleonora Meleti
Thank you very much for these very useful presentations that have started off the debate. I'm quite disturbed by this. In our job, we're really chasing after the problem after it's already spread. So we're trying to find a way to tackle the problem after we've let it happen. Deepfakes become viral first, and then we have to try and stamp them out.
The question I have is how can platforms act preventatively? How can they stop something from going viral rather than us chasing after the image after it's already become viral? I also have an observation to make about hate speech. You say that platforms have a responsibility to identify abusive material. Can users make reports and protect themselves in the same way as they can block the person who is dealing out that hate speech?
We're not really seeing that effort being made by the platforms. You just have to look at the profiles of the people who are around the table here. There's real hate speech out there, and we're not seeing any efforts to restrict hate speech. When we show comments that refer to politically sensitive issues, which tend to raise red flags for these platforms, there we see the is. Is it perhaps that these platforms don't want to comply?
We have rules in Europe. We have laws, but we can't enforce them, and we can't cooperate with the platforms to get these effectively enforced. Thank you very much.
Alessandra Moretti
Thank you, and thank you, madam chair, for organizing this event to look at it in detail. I'd like to thank all the colleagues who have spoken. The Grock case is a turning point, and maybe we don't understand it fully just yet because this is illegal material that's being circulated online. So violence being done further upstream because you have these images which are being created artificially, and it's no longer just a matter of stealing a photo and then sharing it on the Internet. You just have to ask AI to do it for you.
This really changes everything. So it means every woman, every girl, and every minor underage girl is now going to be within reach of these mobile phones. It's not just about violating privacy. It's not just about a lack of consent. No.
This really is the appropriation of the body and the digital identity of people. And then there's another aspect here, and perhaps it's even more worrying, and this concerns young people. This technology is creating a new kind of socializing violence, and these young people don't see themselves as having committed some sort of crime because it's done through an app, it's done quickly, it's done anonymously, and it has been without consequence thus far. So there's a risk here that this act is being normalized, but it really is a serious kind of gender-based violence. So that's why we need to look at a response.
And just using a suppressant response is not enough. We need to look at tech and how respect is included. Now we're always lagging behind. However, we need to be able to anticipate these phenomena because we've got all of these experts. They're telling us what is happening.
They're telling us the direction it's headed. So I think that just trying to ban this is not enough. We need to take action when it comes to education, when it comes to respect, and also when it comes to our digital usage. And we've got the initiative from the commission to step up the responsibility and liability on the side of the platforms. I think that's going in the right direction.
The harm caused at digital level is actually institutional, as we've heard before, and we're not able to protect the victims, but that should be our duty. So I think that we need to work hard on education and prevention, and even access to the technology. So that's why I've got some questions for the rapporteur. Which educational tools do you think we can use in schools to make people understand that creating the image is an act of violence, and it's not the creative use of artificial intelligence? And then another question, is it realistic to think of having some kind of limits to access for young people and to those who are most sensitive to AI and how it's used on social media?
Or is there a risk here that we would just see the phenomenon move? And then my third and last question is, how can we take action upstream? So on the models and on the companies who are developing those tools to make sure that there is prevention included in the design, and it's not just one that we're looking at in terms of the victims.
Lina Gálvez
Sandra, I don't see anyone from Patriot, so we move to ACL, Laurence Rochus, the floor is here for two minutes and a half.
Laurence Rochus
Thank you very much, Madam Chairman. I was going to say something else, but in fact, I want to support the colleagues from the EPP and the S and D who've said much of what I wanted to say. You need a technical solution and a legal solution, but we also need to look at the causes and the prevention. Unfortunately, there will always be loopholes if we don't attack the cause. It's the parents who are the first people to be concerned.
Now I noticed on the slides that it was young people aged 10 to 13 who were most at risk. And I have two daughters aged 25 and 16, and they didn't get smartphones until they were 17, 18. And this is a duty that parents have. We have a duty to inform parents who don't understand the risks of social media. It's important for parents to play their role in the prevention and education towards their own children.
And you may or may not know that, in Australia, there's a platform called Coco, which still allows the distribution of offensive content. We have an agreement with Australia. Australia is a partner of the European Union, and the least it can do is respect the DSA and protect children.
Lina Gálvez
Renew, Abid Al Salani. Please, the floor is yours for two minutes and a half.
Abir Al-Sahlani
Yes. Thank you, madam chair. Thank you, esteemed guests, for your very valuable information and input. And, it goes without saying almost that what we all have heard today is just another proof that gender based violence is no longer confined to physical space and the physical world. It has escalated and scaled, and it has accelerated, and it has gone digital.
Platforms have created a space where abuse is faster, wider, and harder to control, and where women and children are still the primary targets and thus the primary victims. And now AI has taken it one step further. Tools like Grok are presented as honest, unfiltered, different. Let's be very clear about what that means in practice. It means fewer boundaries, fewer safeguards, more harm, because this is where the real crisis begins.
We are witnessing the industrialization of abuse as AI is now being used to mass produce nonconsensual sexualized images that are targeting real identifiable people. This is not innovation, but it is its righteous name is exploitation at a scale. And, yes, it is disgusting, but more importantly, it is dangerous. It is a direct attack on safety, dignity, and fundamental rights, particularly for women and children. It is also an attack on democracy because when women are silenced, intimidated, and exposed in this way, they are pushed out of public life.
And our democracy is dependent on the input of half of its population. Because when women are pushed outside of politics, democracy is no longer to be called democracy. This is not accidental, but it is structural. As a woman in politics, I know visibility comes with a cost, but that cost cannot be my body nor my daughter's bodies, and it cannot be passed on to the next generation. We are already behind.
The DSA was necessary, but it's not enough because tech companies are not self regulating. They are outpacing us. So we need to act fast and decisively. Ban new defier systems, enforce consequences, listen to their victims, and build policy around their reality, not theories that are dressed as safeguards. And most importantly, stop treating this issue as a side issue.
This is not a niche problem, but a systematic threat to our democracy. If platforms feel no shame in violating fundamental rights, why do I feel the sense that we are ashamed of holding them accountable? The time for action is now. AI should be serving humanity as a whole, not abusing and violating half of its population. Thank you.
Lina Gálvez
Thank you. I don't see anyone from the Greens, so we move to the left. Hannah Gedda, who's at Contura. Yeah. Missus, please.
Elena Kountoura
Of women and children, unfortunately, find themselves faced with the nightmare of having fake images and videos of them circulating online showing them naked. This is material made by the use of AI in order to intimidate or blackmail them. Of course, these deep fakes, which were a form of violence, in our report, violence online, which is an ownership opinion, has become familiar, but this kind of digital rape continues to exist with no consent on the part of the victims and no protection for them either. You don't need any technical knowledge in order to produce these images. So even the victims and the perpetrators could be minors.
And we've seen how hard it is to control these kinds of illicit behaviors. So, apart from the attack on the dignity of these victims and the attack on their sense of safety, they are faced with anonymous perpetrators as we've raised in all of our speeches. These platforms have specific responsibilities. They say that there has to be a reporting mechanism for illicit material, which will then restrict access to that material. But when does that happen?
It only happens much later on. And what do we say in the European Parliament? We say that we need to have a legal way to control these platforms with regard to AI produced content and fast response to this abusive material, with the protection of victims included in that process. So with your experience, how do you think that these platforms could be forced to comply? How can we control their compliance?
Is there a way to organize a rapid response or a button, a reporting button that would act immediately in a couple of minutes, so that this illicit content could be taken down immediately. Because if you leave it two days or three days later, the damage is done. And lastly, if these platforms refuse to comply or fail to comply, how can we not only get them to protect victims, but also compensate them for the harm?
Lina Gálvez
Thank you. I don't see anyone from ESN. So I would like to know who other members would like to take the floor. Hannah? Yeah.
Yeah. And, yeah. And okay. So, we start, one minute each, more or less. So Gabriela Gersini from the EPP.
Gabriela Gersini
The floor is yours. Thank you, madam chair. I will speak in Hungarian.
Speaker 28
Thank you. One minute is a challenge. And young women are hit by this problem. And this could be a problem, for example, for my daughter when she uses the Internet. And we haven't seen it happen yet.
But as a Hungarian, I'd like to add that offline there are victims too, and these are children who are growing up in state care or in institutions for young people. And we've seen some situations where there has been sexual violence, and that's a particular risk for those children. I don't know if you have statistics about those groups either. And then another point I'd like to make, something that's very important, and this is something that has already happened. And you have children who are in a spiral of trauma and end up thinking about suicide.
And of course, you need to have people who are trained and qualified, who are available to help support those teenagers and young people, and who can mitigate the conditions. And I'd like to echo what a colleague said. Education is very important. Sharing information between member states is also very important, and it would be good to share good practices as well. That would be very helpful.
Thank you.
Speaker 29
Thank you. Now from S and D, Helen Fritzsson, please. The floor is yours.
Heléne Fritzon
Thank you. Thank you, Cher, and thanks to the experts and rapporteurs. This is a super important debate. To protect children, girls, and women from gender based violence is really a critical issue for us in the European Parliament. We can see these sexual images on the platforms every day.
They are produced when we are sitting here. Yes. Now we know that girls 10 to 14 years old, they are there just now. So for me, it's clear what's illegal offline must be illegal also online. And for this, we need to see that the platforms have to take more and stronger responsibility.
The efforts have not been made yet on their side. And we know that the parents are worried and confused, and the teachers want to have the responsibility to teach and to inform the children. So we need both education and prevention. It must be a part of our daily life. And, of course, we need a ban, and we need age limits for different platforms.
And, of course, also so-called identity protection against deepfakes. So my question, and I can feel we have it together in the room, what's the most important step, the most important tool, our regulation, to put further for us in the European Parliament to reach a stronger legislation and regulation? Thank you.
Thank you. Hannah Gedding from the left.
Speaker 34
Thank you so much, and thank you all for these very valuable insights. I mean, you have given us some horrible numbers today, and I know as well that the European Commission states that pornographic material accounts for about 98% of the deepfakes. And, we know that children are especially exposed to greater risks. And when it comes to grok, we know as well that grok has generated Holocaust denial content. And these are actually companies driven by profit.
Speaker 33
And I think as long as it does not cost them enough to violate our dignity, we have a problem attacking the problem. But, of course, we need to continue putting things into action. We need more pressure on the tech companies and on X to ensure increased industry efforts. And we now need, of course, as has been said today, better to better implement the artificial intelligence act and the DSA. But I'm also interested in what has been said about the preventive measures.
Speaker 35
It has been mentioned preventive systems, the duty of care, etc. The UN Women Agency stresses, for example, investments in digital literacy, especially for young people, and cultural change programs that challenge toxic online communities, including the growing manosphere that we have been discussing in this committee as possible preventive measures. So I would like to hear from you. Do you believe that this would be an additional effective, preventive measure against AI gender-based violence, or what kind of preventive measures would you like to see? Thank you.
Lina Gálvez
Thank you. And now from S and D Libbey Committee, Bergesse Paul. The floor is yours.
Birgit Sippel
Thank you very much, colleagues. To be very honest with you, I'm sick and tired of following this debate. Because, yes, we have a big problem with pictures of people, mainly young girls, but also adult women that are sexually abused, exploited without their consent, and we see it every day. And I think using our bodies, our faces for things that we never agreed to simply is a crime. And I wonder why we are talking about, oh, we need better education.
Yes. We do. And I'm sick and tired to hear we need to have more awareness raising. Yes. But the underlying problem is not our children.
The underlying problem is not even AI. It's the way it is used. And I think at the beginning of our today's meeting, it was clearly mentioned, this is not happening just by chance. It's intentional. There are people producing AI systems.
There are providers. There are platforms who are pushing it because every click, and we heard the numbers, how fast these pictures are spreading around, and every click is creating profit for the platforms, for the providers, for those who are preparing these AI systems. And I think that's something that we have to react to. And, obviously, one of our guest speakers talked about that we set a global standard with GDPR, DSA, and AI Act, but, obviously, the providers are not reacting. In the last month, we had several problem examples where platforms were informed about what crazy, unacceptable, illegal things are happening on their platforms, and they were not reacting.
They try to find excuses. They don't react. So what kind of harsher legislation are we looking for? I think if platforms are not willing to follow legislation to respect our rules, to respect the life of people, the only language they understand is keep them away for a while from our market, because it's profit they are interested in. That's the only answer.
Everything else will make life harder for everybody, but those who are creating these things today will not stop doing it. Thank you.
Lina Gálvez
Thank you. And now I will also make three questions. My first question is very much in this line because we have been talking here about this preventive way. But if it is that possible, have any mind that probably the problem is in the business model itself? I mean, and we know that violent content or hate content normally has more likes than other types of content.
So that means more time staying on the platform. That means that business will have more chances to sell those goods or services in line because people will stay more time and also to take their data in order to trade that data. So it is on the basis of the business model. So my first question, it is very much similar to the Virgil Sipu was doing. It is possible with this business model to speak about prevention, to go on this prevention way.
The second one is something related to education and something you have said that almost also the boys that are producing these fakes are victims. And this is something we really need to highlight as well, that they are victims also of the system. They are victims of this systemic way. We talk a lot about education, but education is not only what happens in school. It's not even only what happened at home.
It is everywhere, and it is mainly in media. It is also on this social media. It's on pairs. So even if you do a lot with your own daughter or son, could be others that are influencing them much more than parents. But still, I think we do not have digital literacy.
And for a long time, not only now with all those problems, we have not taught children and adults on reading images. I mean, as a feminist, I have read a lot on how women are treated on fashion and publicity and how we are it's just bodies without hats and all these little things. And we are not, we're we have not learned how to read those images. So this is entering in our mind in a very natural way, and we are not teaching. So still, there is still possibility, I think, to improve.
And another thing is how it is realistic to talk about a huge engagement of parents when we are using that all the time too. And we are suffering, most of us, of time poverty, so not having enough real time to spend with our children. And the third question, it is on international scenario and international cooperation. We know that a lot of these harmful practices are done outside our borders. So I was talking with our Korean colleagues before, say most of these perpetrators are coming from outside or are done outside Korea.
In the European Union it is the same because at the end, the digital world has no borders. So this is very important for an important issue here, that it is accountability. So I have two questions regarding that. First, do you think sanctions are possible within the Digital Services Act up to 6% of worldwide revenues, not just the revenues done in Europe, but worldwide revenues. It is a good tool also, not only for the European Union, but also for making these big companies accountable worldwide.
Because if you take 6% of their revenues, it's not a bad issue. So you have to be very brave. I mean, I would like to see if European Commission will go that far, regarding all the pressures that we are having as European Union on not developing all our digital regulation. And, but I would like to know if that will be that possibility, I mean, to go really too far, in order for this big platform to react. And the second question regarding this international is to what extent do you think that after Trump, or the arrival of Trump into power, it is aggressively going against any intervention on the digital sphere, very anti-feminist, very anti-gender, and especially backing this kind of Far West model of this big tech with all the tech bro bros, including his government, so on and so forth, is more difficult than, let's say, one year and a half ago really to move on this international arena.
And now you have the floor back. So starting by you, do you have, yeah, five minutes each or even a bit more to react because we have time? Thank you.
Thank you for the five minutes. I really do appreciate that.
Suvi Uski
And then more. We have time. I mean, if you want to take 7, it's fine too.
Very good. And I'm impressed by the discussion we're having in this room. It really gives encouragement also to my work and my team's work back in The Nordics. Usually, when I meet with professional educators, I use this very simple kind of example to kind of put everyone on the same page about what we are tackling here with. So, usually I tell that if we look at adult average social media use per year, it is 870 hours per year.
So, it is like two hours 23 minutes per day. And it is not too bad. Like ten years ago, we had these screen time recommendations that, okay, two hours is something that you can do in a healthy manner. But when we put the youth and teenage social media usage on the side of that 870 hours per year, what we get is 1820 hours per year. So, it's 1000 hours more per year if you are a younger person.
Okay. I come from social psychology. I'm like, okay, what are these young people getting from social media? Like, the benefits if they are spending 1,000 hours per year more. And the fact is that if we look at the research of any kind, the recent ones, there are no benefits really.
It's only the negative sides. And then my second question is, okay, as like you know, I meet with professional educators. How can you entertain a young person for seven to twelve hours per day to hang out with a smartphone? What do you show the children so that they get sticky with the device? And that's where we come to this picture as the topic we are discussing here today, that the content that is being shown is very algorithmically unethical.
And this change in the numbers of usage in social media, it has come into place after COVID. So after 2020, and that's not because our young people were somehow ruined or something happened in their culture. No. It's not that our young people's or children's fault that they're spending so much time on social media. It's with the algorithms.
And it's with the algorithms that came with one very popular platform, which were super unethical, which we didn't use here in the Western countries, really. And now slowly but surely what has happened is that these unethical algorithms have taken place in all of these super popular social media platforms, which actually means that we, everyone also here in this room, are spending more and more time with different types of content. And most likely, there is harmful content mixed with the good stuff to keep you on the platform. And when we come to the regulation part or when we come to the part where we are okay, what's secure for this ache? We need to ask, what are these platforms doing?
What, why aren't they being more ethical? Why aren't they serving our health or well-being? And the question or the answer is that they are competing with each other. And that's the business model that you were kind of raising here, and that's that they are actually competing for the time of the user. And, unfortunately, I've been working with parents, I've been working with educators in all levels in our societies in The Nordics, and I just as a social psychologist, I know that these algorithms have hacked our psychology.
So, as an adult, I can't resist my device. So how could I expect that from a child whose psychological development is not fully in completion? And that means that we are tackling something that even parents can really take a hold onto. And also taking into consideration that five years ago, the instructions from the school, even from the kindergarten, were, like, more positivistic towards technology, that, yeah, please bring the iPads to the education. And that was a good idea back then, but we didn't know that this unethical algorithm thing will kinda change everything we do with our daily digital devices.
So, that's the one thing. And then also about education, if we talk about media literacy, it's not enough, unfortunately. I'm doing that every day. We have our own platform where we educate kids of all ages. And we also have their quiz, how to make a difference between fake and real images.
I couldn't get it right. You know? I didn't know which was fake and which wasn't. And I am professional in this area and I don't know. So how can I teach my kids or in our countries that, okay, this is the way how you will recognize and this is how you will tackle this?
You can't. It's impossible. And same goes with these deepfakes today. And then also, I think what we need to think of, like, in terms of technology, it's not neutral. The beholders of technology really decide what they do, and they could have a lot more safer platforms.
And it's very unfortunate that we need to hear to tell them what to do. Like, it's not really good use of our time and to put us into this position that we should be here designing their technology if they can't do it themselves because they can. And they have all the access to change the features and build the safety within. It has been there. If we had the direction when we go twelve years back in the social media history.
We used to have this, moderation was improving. We had more, like, safeguard trails. It was, like, a nice development, but then that was taken away, unfortunately, and that's something that we witness at our work every single day. Yeah. Thank you.
Lina Gálvez
Thank you. Thank you very much. And now, I would like to pass the floor back to Segi Akiwawa, who was with us online. A good thing of the smartphone is I have already downloaded your briefing. You see?
So there are also some good things. But so, please, the floor is yours.
Segi Akiwawa
And that tech means that I can be here in the UK looking after my puppy and still engage with you guys in Brussels. Right? Like, there are beauties of tech. There is some points I wanna respond to. I'll take them in turn.
The first one on engagement metrics. I think we have to be really careful to not buy into the language that the big tech companies have monopolized. Engagement metrics is a decision. It's a decision to set those engagement metrics on things like shame, outrage, envy, hate. You can set engagement metrics for love, for joy, for hope.
We saw that during the pandemic when we were seeing dance challenges. We see that when we in explore pages, on our social media feeds, we see content around dogs and puppies. There is a way of setting engagement metrics that does engage us as human beings, as people with interest that isn't about the worst of us and isn't about bringing up unhealthy behaviors in us, which brings me to my point about some new research that I'm working on with a telecoms company in the UK. Emerging research shows, after interviewing 6,000 people, that one third of respondents feel able or competent in being intentional with their social media platforms, with their device. So there is a gap between social media use, screen time use, and intentionality.
I don't think it's a blanket screen time numbers equals bad, because we now know that we can code on social media. We can connect with communities on social media, particularly for communities like mine where inequality means not being able to get access. Let's remember the good times of social media where it democratized information, and it reduced the hierarchy between celebrities and writers, and it provided possibility models for young people who would never get the access that mostly privileged people would do. So I don't think screen time hours equate to bad. I think what we need to move into is the question around education: what is our intentional use of social media?
What's our intentional relationship with platforms? And that does start by education. That does start with giving investing back into critical thinking, investing back into emotional resilience, being able to regulate and calm, and pause and say, something has sparked me in an emotional way. I need to take a minute and breathe before reacting. Because we know that education is gonna be our only way out of this, because banning is proving already to show unintended consequences.
Australia have enacted the social media ban for under sixteens, and emerging evidence just last week from Chicago show that only one third were respecting the ban, that at least two thirds of young people need to be adhering to the ban for it to be effective. So we know that prohibition for social media is not going to be the answer here. It's education. So how do we start doing that in education? I completely agree with you.
It can't just be around parents and families. Because, again, families like mine where it was a single mother who's working two jobs, where are you gonna ask them to have the time? And you're gonna keep relying on NGOs who are having to deliver services on a breadcrumbs, on a shoestring budget from philanthropists who are not having a steady strategy. It's not a sustainable model. The funding of education provision, which should come through libraries, it should come through hairdressers, it should come through the barber shops where there is information available to talk about social media, real healthy relationships, and instilling this village approach is through the taxation of tech companies.
I do think sanctions and taxation work. We've seen this in other industries, like oil. But what we need to make sure is that this devil is in the detail, that we ring fence a portion of that sanction to go back into mitigating the harms that was spoken about by a lot of members in this committee, that we need to make sure that a portion of the worldwide sanctions is going back to reinvesting in education, enforcement, and readdressing harms. Because as somebody who was a victim of online harassment and abuse, speaking at the European Parliament that went viral, I was now facing hate speech and harassment and death threats from outside of the UK. This is a borderless crime, and so we need to make sure that we are investing in the resources and the capacity to enforce this.
So to end, I will say that I was glad to receive an email from the omnibus group that was talking about the specifics that could be added to the wording around this. So we don't have this loophole situation where we where it or it's too or the wording is too limiting. My suggestion is that we have a practical addition to the wording that says that classification and escalation frameworks across platforms and institutions should be shared, and that harm is not just prohibited in principle, but is consistently recognized and acted upon. So what that looks like is not waiting for it to be visible harm, but knowing, as many members said in this committee, that the fact that there is psychological impact, economical impact, that alone is enough to take down the content because we're then moving regulation away from reacting to content and bad actors and bad business models to healthy standards and approaches that we want to be seeing. Frictions that are good for the betterment of society, if it does mean at the first instance, it's a bit uncomfortable.
And this is probably where I do want to end. A lot of intervention is needed by tech companies, by policymakers. I wish my own UK government was as passionate as a lot of you members are about regulating this. Because as you've also seen in the UK, we've allowed must to provide energy, have energy contracts. So, you know, anyway, Brexit.
Alright? So culturally, we need to have an understanding that this is not just a tech problem, but a cultural problem. And this was evidenced with the inaction of age assurance around pornography. Last year, we saw a backlash to simple friction of having to verify your age so that we stopped young boys as early as 11 accessing some of the most violent, disgusting pornographic content. We, as a culture and as a society, have to also start saying we're okay with some change, with some friction to our tech, so that everyone is having a healthy relationship with technology.
Speaker 36
Thank you. Thank you very much. And now, David Rachel, the floor is yours. Thank you.
David Rangel
You for the floor and for the very good questions. I tried to address each MEP's question, but there are so many, and I tried to cluster them. I think, as was mentioned, for example, the sharing of nonconsensual images created, manipulated or not, is such an obvious attack on the dignity of women that I think it is also a very good test to the legislation that we have. And, obviously, as all of you agreed, is that existing legislation needs to apply. And to see how well it works, needs, of course, several steps.
Speaker 7
One of them is, unfortunately, very often, we still need to prove a lot of things, and that's where more knowledge is needed and the data collection on where the problems lie. And that is actually not too easy, because on the one hand, it requires quite some resources to actually prove certain harms of online platforms. And secondly, and that's another problem, is that access to relevant data off the platforms is a challenge because the platforms are not easily allowing them to be investigated. That's linked to one of the cases under the Digital Services Act, provides a provision for researchers to access the data of online platforms in order to investigate systemic risks. And systemic risks, I think that's also the term that speaks to challenges that could then be analyzed in a way how these new tools may normalize certain problems as you mentioned to really understand where this goes in the wrong direction.
And this way, also anticipate certain problems. But unfortunately, and me personally, it sometimes reminds me, like, what happened in the seventies with the industry. You had to collect a lot of data to prove that it causes cancer. Maybe we're in a similar situation with online platforms to still having to show that it harms, even though in certain aspects it is very obvious. So knowledge collection and production to prove the cases in order to apply the legislation, I think, is an important aspect.
Secondly, to prevent causes, I hope that there is also the idea of sustainable businesses. So we have some data from our antisemitism survey where we asked Jewish people in the EU about their experiences of antisemitism online. And we clearly see that those encountering antisemitism also stop using certain apps in practice. So it also has an effect on the business case. If it's strong enough, we still need to see, but that's obviously something we hope that works as well.
The third point that was raised, I would like to react to, is access limitations. So the data I presented concerns people, children encountering certain content online. We also asked about social media restrictions. As you know, this is a political topic at the moment, and it is, in fact, the majority of people in the EU are in favor of social media restrictions. Although there's no agreement on what age it should be.
Interestingly, in our data, we also looked into the 15 to 17 years olds that were covered in the service, so teenagers. And in fact, also there, the majority agrees that social media restriction should be. Usually, they would say below their age, so that's to be acknowledged. But at the same time, it also shows that teenagers are aware that it is dangerous to use certain tools. However, having that said, we know that restriction is not the single solution and not the solution potentially because children also have rights, and child rights don't only mean that you have to protect children, but also to empower children.
So just cutting down the access as such is not the solution. And as you know, the European Commission has an expert panel that advises President von der Leyen on this topic at the moment. So this is being looked into. There's also the guidelines for the protection of minors under article 28 of the Digital Services Act, that platforms have to adhere to. That also includes risk assessments for minors in using the tools online.
Yeah. Because the restriction as such is not a solution, and it needs to be a balanced way on how to find this, for example, in combination with the importance of education that we discussed and how also teenagers and young people use social media. But we also heard, like, there are other topics. It's not always about illegal, sometimes content is harmful, and it's difficult to address. We currently run a project on child safety online where we look into topics like self-harm, but also the topic of boys was mentioned.
Toxic masculinity, is there an area? And we so we speak to digital services coordinators and see, like, what could be potential solutions to also address these risks, on social media. Yeah. There was the point I would like to suggest one step, but I don't think there is just one step we have to take. We have to take a lot of steps.
We have several different tools. If there's no compliance, of course, sanctions are foreseen in the legislation. We know from data protection law that it can help if it's needed in practice. Yes. And I would not wanna comment on the politicization of the enforcement of laws from your good question.
But just to remind, maybe as a closing that, I mean, the Charter of fundamental rights is EU primary law. It enshrines the basic rules of our society. And I think, if we see that the laws don't protect those rights, then, of course, further action is needed.
Lina Gálvez
Thank you. Thank you. Thank you very much to our experts. Also, thank you to the members. Probably, this is one of the topics we can find more agreement within this committee, in the sense that it is a scarring topic, and it is linked to systemic risk.
And also, we all think that we are lagging behind in our answers. Or even if we put answers, everything is going faster than what we are doing. So I would like to just express my sincere appreciation to all the insightful, valuable contributions. One thing is clear. This intersection of artificial intelligence and gender based violence is no longer a future concern.
We are, as I said before, lying behind. It's already here. And the case of Brock has shown us how quickly powerful technologies can be misused or even misdesigned. Not only misused, but also misdesigned or designed for some purposes that are against our fundamental rights and how easily can amplify harm, particularly against women and children through the creation of circulation of sexualized or pornographic, exploitative images. So we have been discussing, as you have here, the technology, it is not neutral.
And, but their use is even less neutral. And, obviously, what is clearly behind, it is a patriarchal society that undervalues women and reacts also in a violent way if we are not in the places the society wants us to be when we are performing a role. We are not expected to be. So this is not just a technical issue. So for us, it is a very core issue in our committee, and because it can also reflect and reinforce assisted inequalities, biases, and forms of violence that once more affect disproportionately women and girls.
Artificial intelligence systems do not operate in a vacuum. They are shaped by data, values and safeguards that are chosen to implement, including this specific business model. What we have here today underscores the urgency of action on multiple fronts. As you were saying at the very end, it's not the only way to answer. And we have an important legal framework, but we really need to implement it in the right way and go further if necessary, even if banning, it is obviously not always the way or the only way. We have really to use all the tools we have in our hands.
And at the end, if I have to highlight one sentence, it is that we need to be brave. And unless we are brave and we act united, we will not be able to stop this problem that, you were mentioning, has affected specifically a generation because I have a daughter, 15 years old. And since she was in the kindergarten, it was like to use a technological device. It will be good for her development. And now we have seen that this is not the case.
So we are taking measures, but there is a generation. I will not say there has been ruin, but it has suffered the most that we are seeing now also in the anti gender, for instance, that it is and all the conspiracy and all the antidemocratic that it is going on, especially on this very young generation. So we have a very serious problem, and I hope we will be able to tackle it before we did with the tobacco industry. But as it happened with the tobacco industry, there were a lot of interests behind and a lot of money making society believe what was not true. So a lie.
And now it's probably even the same because there is even more money. So, thank you. Thank you very much. Thank you also to the interpreters, and we still have ten minutes before coordinators. So, thank you.
Thank you very much, and we will definitely keep discussing this very important topic. Thanks a lot.
Ask anything about this video — I analyse its transcript to answer.