Women in an artificially enhanced information space Countering gendered AI‑generated information manipulation
Summary
While new AI technologies promise to advance productivity and growth, adverse consequences of the fast rollout are already materialising. In particular, negative impacts on gender equality are visible across societies, across policy areas and across the world, often accelerating reversals of previous positive trends. AI is already amplifying gender bias in the information space, as well as enabling and facilitating malign creativity, including behavioural manipulation, global influence campaigns and the production of non‑consensual intimate imagery (NCII), primarily harming women and children.
AI‑accelerated pressures on women in the online information space can cause severe harm to individuals and can hamper their democratic participation. They go hand in hand with a global backlash against gender equality and women's rights. Gendered disinformation – both a strategy to silence women and gender‑diverse groups, and at the same time constituting a form of online gender‑based violence – is seen as a warning sign of erosion of democratic principles and institutions more broadly and has been recognised as a security threat. Women in visible political and public roles, not least women politicians seeking higher office, but also journalists, experts, human rights defenders and prominent public officials, face higher volumes of sexualised attacks designed to undermine their credibility and deter them from participation.
In the European Union, equality between men and women is a fundamental right, and a founding value enshrined in the Treaties. Over the past decades, the EU has worked to increase gender equality at home and used its global influence to support women's rights abroad. At the same time, its evolving digital regulatory framework aims to make the digital space safe for all, including protecting the rights of women online.
Key concepts
Disinformation: Verifiably false or misleading information that is created, presented and disseminated for economic gain or to intentionally deceive the public, and may cause public harm (COM(2018) 236).
Foreign information manipulation and interference (FIMI): A term introduced by the European External Action Service (EEAS) in 2021, FIMI is described as a 'mostly non‑illegal pattern of behaviour that threatens or has the potential to negatively impact values, procedures and political processes and which is manipulative in character, conducted in an intentional and coordinated manner by state or non‑state actors, including their proxies inside and outside of their own territory'.
Gendered/identity‑based disinformation (IBD): The EEAS's 2024 practical guide to detect and analyse identity-based disinformation/FIMI refers to IBD‑focused FIMI as 'a non‑illegal pattern of behaviour that threatens and causes direct or indirect harm to individuals and groups based on their gender, orientation, race or ethnicity, affects values and disrupts the political process. Such activity is manipulative, intentional, and coordinated, and it is conducted by state or non‑state actors, including their proxies inside or outside of their territory.' EEAS calls for IBD in the context of FIMI 'to be recognised as a serious threat', and for open source intelligence practitioners to 'maintain an identity‑focused lens' in their work.
The UN Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression noted in 2023 that the 'dual nature' of gendered disinformation – on the one hand, a strategy to silence women and gender‑diverse voices, and on the other a form of online gender‑based violence – makes both the definition of the problem and the search for solutions more complex and challenging'.
Non‑consensual intimate imagery (NCII): Sexually explicit photo or video of a naked or partially naked person, showing their genitals, engaging in sexual activity or poses, or wearing underwear in compromising positions which is captured and/or shared on‑ or offline without the explicit consent of the person in question (source: the EU's terminology management system IATE). NCII can take the form of non‑consensual synthetic intimate imagery or non‑consensual sexualising deepfakes, which emphasises the harm actively inflicted on the targets, as opposed to the more passive notion of non‑consensual sexualised deepfakes (Pawelec and Łabuz, 2025).
According to experts, the notions of 'deep fake porn' or 'revenge porn' are problematic because they imply consent. In addition, the latter incorrectly suggests that the depicted person has played a role in causing harm that deserves revenge (Morgan, Farid and Nightingale, 2026).
Introduction
While new AI technologies promise to carry the potential to advance gender equality and women's rights as well as increase scientific advances and productivity at large, adverse consequences of the fast rollout are already materialising. In particular, negative impacts on gender equality are visible across societies, across policy areas and across the world, often accelerating reversals of previous positive trends. AI is already enabling and accelerating malign creativity, including behavioural manipulation, global influence campaigns and the production of non‑consensual synthetic intimate imagery or sexualising deepfakes, primarily harming women and children. In a gendered attention economy, young people – notably men and boys – are algorithmically fed sexualised content and misogyny, exacerbating biases and exposing them to more extreme content.
Growing AI‑facilitated pressures on women in the online information space go hand in hand with a global backlash against gender equality and women's rights. A rise in violence against women online overlaps with gendered disinformation using affective polarisation, increasing animosity between people, widening the gap between positive emotions towards the ingroup and negative feelings about the outgroup (Iyengar and Westwood, 2015). Di Meco (2023) sees gendered disinformation as 'an early warning system for both backsliding on women's rights and the erosion of democratic principles and institutions', which can – when deployed by foreign malign actors to exploit divisions in society – constitute a (national) security threat.
Research shows that those who challenge – or are perceived as challenging – existing gender hierarchies are more likely to be targeted. Visible women in prominent political and public roles, not least women politicians seeking higher office, but also journalists,1 academics, human rights defenders and prominent public officials, face higher volumes of sexualised attacks designed to undermine their credibility and deter them from participation (DiMeco & Wilfor, 2021, Posetti et al., 2021). Public attacks on women role models – enabled and amplified by the digital infrastructure including online platforms and AI tools – can signal directly or indirectly to other women and girls that participation in public life comes with (too) high risks.
According to the 2023 report of the Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression (A/78/288), gendered disinformation has multiple aims: 'portraying women as weak, incompetent and sexualized objects, incapable of leadership; driving women and gender nonconforming persons out of public spaces and places of power; and silencing those who do not comply with gender norms'. Gendered disinformation attacks 'not only individuals but also their collective struggles by seeking to delegitimize feminism and gender rights'. The overall objective is to undermine human rights, gender equality, sustainable development and democracy.
Gendered disinformation
Political disinformation attacks
The number of documented cases of foreign interference in democratic elections worldwide continues to grow. This includes gendered disinformation campaigns targeting female candidates in US presidential elections, such as Hillary Clinton (2016) and Kamala Harris (2020 and 2024), explicitly focusing on their sex. In Europe, prominent political targets of AI‑enabled gendered information campaigns include former German Foreign Minister Annalena Baerbock, Moldova's President Maia Sandu, Italian Prime Minister Giorgia Meloni and former Finnish Prime Minister Sanna Marin, among many others. Prominent women in other influential positions, such as judges and government officials, rights advocates and journalists, are also targeted with disinformation, often combined with online and offline (political) violence. In the context of the European Parliament elections in 2024, the Global Disinformation Index observed a rise in online gendered abuse and disinformation about female EU leaders, including misogynist narratives and slurs to portray female political leaders or candidates as less qualified than their male counterparts.
Key gender bias narratives
Kristina Wilfore (quoted by Di Meco, 2023) notes that 'gendered disinformation campaigns build on, and are rooted in, deeply set misogynistic frameworks and gender biases that portray masculine characteristics as those fit for leadership while painting women leaders as inherently untrustworthy (insinuating a woman is dishonest or not trustable is a tried and true attack), unqualified (one of the biggest barriers women face when seeking office), unintelligent (tropes about women as dumb and unfit for the job are a prominent feature of gendered disinformation, made worse with objectifying sexualized content), and unlikable (which for women can be the death knell of their campaign)'.
In a 2024 limited survey of 15 US National Democratic Institute field offices (unpublished report cited by the Centre for Information Resilience), 100 % of respondents said that gendered disinformation was utilised against individual women in recent elections. Illustrating the connection between individual gendered disinformation attacks and the broader pressure on rights, 87 % of respondents linked gendered disinformation campaigns to strategies to support democratic backsliding or the rollback of women's rights. Women democratic leaders are often some of the earliest targets of such narratives. In interviews with women targeted by Russian, Iranian, or Chinese state media, researchers found that gendered tropes formed the basis of disinformation and smear campaigns against them (Jankowicz et al., 2021). Part of the strategy is to 'humiliate, shame, and objectify women, especially women who have the temerity to speak out' (Jankowicz, 2023).
Sexualising deepfakes: Accelerating AI uptake and revenue?
Generative AI tools are rapidly transforming the information space, accelerating the creation and spread of mis‑ and disinformation, and eroding the business models of traditional news media and knowledge providers. The most widespread use of deepfake technology appears to be the creation of NCII, also sometimes referred to as sexualising deepfakes. Although NCII is illegal in many countries, tools and incentives to create such content synthetically continue to be readily available: one in three deepfake tools allow users to create sexualising deepfakes. In a September 2026 study, researchers mapped the infrastructure that enables and monetises the spread of such content; they identified 88 publicly available sites hosting such content, with 38 of them dedicated exclusively to synthetic imagery (Morgan, Farid and Nightingale, 2026).
A 2023 report on the 'state of deepfakes' assessed that deepfake videos had increased 550 % online from 2019 to 2023. Another study estimated that the number of deepfakes shared across social media platforms increased by 1 600 % from 500 000 in 2023 to 8 million in 2025 (Khalil, 2025). The 2023 report found that 98 % of all deepfake videos online were sexualising deepfakes, and 99 % of individuals targeted were women. This type of content goes viral easily, feeding into commercial ecosystems of misogyny. Gamification further incentivises this proliferation, for example through 'daily challenges' that motivate users to engage with and create adult AI content, feeding into thriving marketplaces for online misogyny. Such attacks can have multiple motives, including financial, social or ideological motives. Motivation aside, the proliferation and consumption of the abuse is enabled and facilitated by a sophisticated digital infrastructure. In 2023, Graphika found an increase of referral link spam for NCSII services within less than one year, including on Reddit and X.
The UK‑based Institute for Strategic Dialogue (ISD) has identified 'a mature and adaptive gendered deepfakes ecosystem that extends beyond fringe communities into mainstream digital services, including search engines, app stores and social media platforms'. ISD research published in July 2026 showed that over 5.7 million visits to sites and apps with nudifying tools came from mainstream social media platforms, with YouTube (1.82 million), X (1.3 million), Reddit (691 000) and Facebook (464 000) as the main sources.
In a January 2026 investigation, the Tech Transparency Project found 55 apps in the Google Play Store and 47 apps in the Apple App Store that allow users to digitally alter or remove the clothes from women. These apps had been downloaded over 705 million times worldwide, generating US$117 million in revenue, with Google and Apple profiting from the activity of the apps. At the same time, AI companies – which are under significant pressure to generate revenue – are competing fiercely for their share in the booming market for AI‑generated adult content, worth US$2.5 billion in 2025, and expected to grow by 28 % annually until 2028.
Recent reports published by Berlin‑based thinktank Agora (Shultz, September 2026 and October 2026) have demonstrated that women parliamentarians in Europe are far more likely than male parliamentarians to be targeted, echoing similar findings from the US Congress (Li, Shultz and Jankowicz, 2024). Women parliamentarians in visible senior positions are more likely than others to be targeted.
-
In national parliaments in 22 EU countries and across the political spectrum, women parliamentarians are 33 times more likely than men to be depicted in or associated with sexualising deepfakes; 138 women parliamentarians were depicted in or linked to sexualising deepfakes, compared to nine men (or 1 in 14 women, versus 1 in 500 men). Countries with deepfake legislation in place did not show lower exposure, which indicates a lack of enforcement.
-
In the European Parliament, 50 MEPs (46 women and four men) were found depicted in or associated with sexualising deepfakes (equalling 1 in 6 women MEPs across political groups).
Actors and impacts
Strategic state or state‑sponsored actors
A now archived page on the website of the US Department of State (DoS) cites a joint 2023 study by Canada, the European External Action Service (EEAS), Germany, Slovakia, the United Kingdom, and the US, which analysed the tactics used by state actors to sow gendered and other identity‑based disinformation across the world. According to the DoS, 'both foreign state and non‑state actors strategically use gendered disinformation to silence women, discourage online political discourse, and shape perceptions toward gender and the role of women in democracies'. The DoS underlined how foreign state actors – specifically Russia and China – targeted women with gendered disinformation, 'utilizing their media assets, control of the information environment, and sometimes state‑backed troll farms to spread disinformation about women politicians, policy makers, journalists, and activists, and even policies targeted at women'. Russia's weaponisation of gendered narratives in the context of the war in Ukraine is well documented, and at the same time, women have become the key warriors in Ukraine's effort to combat disinformation and strengthen the country's collective cognitive resilience, which creates yet another incentive for pro‑Kremlin actors to discredit them.
In parallel, misogyny as a tool of political repression facilitates digital attacks on (exiled) women activists and thus can be used as a tool of transnational repression. This helps authoritarian regimes to expand their reach and control with a layer of digital transnational repression. Research has found that 'gender‑based digital transnational repression extends forms of authoritarian domination from domestic and "offline" settings into digital space and the transnational field'.
In recent years, gendered disinformation has been recognised as a matter of national security. The 2024 NATO Policy on Women, Peace and Security (WPS) details how strategic rivals and potential adversaries 'promote gendered disinformation to sow division and destabilise our societies', thus challenging allies' 'security, resilience, interests, values and democratic way of life'. However, the Royal United Services Institute (RUSI) warns that the WPS agenda is under renewed political pressure. Gender equality, women's and LGBTQ rights are facing an increasing geopolitical backlash, and gender mainstreaming – ensuring that policy, programming and law prioritise achieving gender equality – can struggle to gain traction when it is treated as politically contentious, or less urgent than immediate security demands.
Anti‑rights pressure
As the Carnegie Endowment for International Peace (CEIP) notes, for decades the United Nations and other international organisations have been key to advancing gender equality globally. In recent years, the same organisations have become strategic arenas for what UN Secretary General Antonio Guterres has called a 'kickback' against women's rights. CEIP asserts that some state actors and advocacy groups are 'resisting new gender equality commitments as well as contesting the language of previously agreed‑upon norms, particularly around sexual and reproductive rights, LGBTQ protections, and references to "gender" more broadly'. In practice, according to reliefweb (the information service of the United Nations Office for the Coordination of Humanitarian Affairs), some governments 'are manipulating multilateral mechanisms to undermine gender equality, and withdrawing from or weakening international treaties and UN bodies, while those defending human rights are experiencing mounting repression and risks'.
Anti‑gender narratives – including from Russia, which decriminalised domestic violence in 2017 – imply that liberal democracies are morally corrupt and decadent, as opposed to so‑called 'traditional values' that have become part of Russia's soft power toolbox. This broadens the scope of existing false allegations of anti‑Russian bias in international institutions, including the European Court of Human Rights (ECHR), the Organisation for the Prohibition of Chemical Weapons (OPCW), the World Anti-Doping Agency, the United Nations, and also the EU. On the other hand, rejecting gender equality is framed by some authoritarian actors, including Russia, as resistance to neo‑colonial domination and to international institutions that are portrayed or perceived as tools for Western influence.
Values‑related aspects of gender equality appear to have become a strategic gateway to geopolitical contestation. In this geopolitical context, gender equality and related human rights frameworks are portrayed as neo‑colonial interference, as reflected in debates over the ratification of the EU–ACP Trade Agreement and the Istanbul Convention (Datta and Velizanin, 2026). Such narratives can not only delegitimise the contested institutions, but also hamper cooperation between liberal democracies and 'Global South' countries. At the same time, the delegitimisation of democratic institutions and bodies – public administrations, governments, health services, scientific bodies, experts – erodes trust in democracy as a system.
A gender gap in trust
In light of the aforementioned impacts on women in the information space, it may not come as a surprise that women tend to trust – and use – AI less than men. Women are 25 % less likely to use AI tools than men. New research contradicts the notion that the gender gap can be ascribed to 'lack of skills, interest or access'. On the contrary, the more knowledge and understanding about AI women gain, the more cautious they tend to become. According to a systemic review of gender and adoption of AI, women mention the lack of transparency and the opaque nature of AI tools as decreasing their level of trust.
Women's hesitancy to use AI may also be linked to a real‑life competency penalty that women face when using AI in the workplace: identical AI use by women and men produces different judgments of the person's integrity and competence; women who use AI are 22 % more likely to have their trustworthiness questioned, as AI use by women is perceived as inability, whereas AI use by men is seen as pragmatic.
Distrust in the tech industry among many women may at least partly be linked to misogynistic content they encounter online. A 2025 survey by Amnesty International showed that 7 in 10 Gen Z social media users have witnessed misogynistic content, with 5 in 10 saying they observed such content on a weekly basis. Almost half of respondents (47 %) blamed statements and actions by tech leaders for worsening the problem. While no specific leaders or statements were named, public comments by influential AI CEOs – for example, that AI is set to disrupt 'highly educated, often female voters', that 'liberal white women' have been 'programmed' to 'hate white people', and that 'the vast increase in welfare beneficiaries and the extension of the franchise to women' made 'capitalist democracy impossible' – could be perceived as mirroring the industry's attitude towards women, thus possibly contributing to eroding trust.
Impact on democracy and democratic resilience
Gendered disinformation can harm individuals and limit their participation in society and democracy at large, with repercussions for democracy and democratic resilience more broadly. Saner et al. (2026) note that gendered disinformation 'intensifies societal divisions, drives dangerous levels of polarization, erodes democratic institutions and trust, and accelerates democratic backsliding'. By doing so, it creates 'national security vulnerabilities by enabling foreign information manipulation and weakening international alliances'. Growing political extremism and polarisation can contribute to a 'self‑reinforcing cycle where [gendered and identity disinformation] both exploits and intensifies societal divisions'. The interplay between foreign and domestic actors makes attribution difficult and potentially contentious, and can further accelerate a vicious cycle of domestic polarisation which can be exacerbated by malign foreign actors.
At the same time, the impact of search engines as information gatekeepers, which use AI to curate political information to their users, shape perceptions of political realities and women politicians. Analysing cross‑national political Google searches, Rohrbach, Makhortyk and Sydorova (2026) found that algorithmic underrepresentation can actively reinforce gender‑biased views of politics. Translated into algorithmic biases, they can 'not only diminish the perceived chances of winning an election for minoritized candidates but also result in voters feeling that their voices matter less'. In this way, 'algorithmic systems like search engines can negatively impact electoral processes'.
Impact on information integrity
Gender bias in large language models (LLMs)
The European Union Agency for Fundamental Rights (FRA) has repeatedly underlined the risk that AI systems based on incomplete or biased data can lead to inaccurate outcomes that infringe on people's fundamental rights, including non‑discrimination. The risks of bias in LLMs are well documented: most LLM companies choose to train their models on sources that are free of charge (or pirated), which can compromise the quality. One example of gender bias in LLMs has been observed in literature reviews, which have been found to amplify a widely recognised phenomenon in academic referencing, in which publications by female scientists generate fewer citations than men's (the 'Matilda effect'). LLMs have been shown to consistently fail to mention female authors in female‑led literatures; claim that men are more influential and more frequently cited, even when this is contradicted by citation counts; and attribute women's work to hallucinated male scholars.
Gender bias in multilingual training data sets
Wikipedia content features prominently in the datasets LLMs are trained on and are consistently made visible in Google search results and overviews. Wikipedia offers a wealth of articles on a wide range of various topics, in some 300 different languages (albeit with varying numbers of articles and editors). Popular content tends to be updated on a regular basis. However, Wikipedia is also prone to not only deliberate manipulation by state and non‑state actors, but also to gender bias, impacting the content and the topics covered. Just 19 % of biographies in Wikipedia are about women, and only 10‑15 % of Wikipedia editors are women. Wikipedia has acknowledged that consistent efforts to improve gender balance have failed to reach its goals. Based on feedback from female Wikipedia editors, reasons why women do not edit Wikipedia include not having enough free time; and the perception of the overall atmosphere as misogynistic and/or sexual in ways they find off‑putting.
Pay gap and credit score bias
Research has identified pay gap as a reliable measure of bias in LLMs. A 2025 study found that, when asked for advice on salary negotiation offers, AI chatbots proposed significantly lower salaries to women compared to men with identical profiles.
Similarly, as AI is increasingly used to assess borrower data, research has found that lending systems still disadvantage women and minorities, thus hampering financial equality: women consistently receive lower credit scores than men, despite exhibiting lower default rates and controlling for extensive credit risk variables. Some research mentions Kenya as an illustrative example of gender bias in fintech lending algorithms, consistently offering women significantly smaller loans than men, equalling a 37 % underfunding penalty for female entrepreneurs.
Countering alleged 'woke' AI or seeking epistemic power?
While such examples result from amplification of existing incomplete or biased data, Korolczuk (2020) has argued that 'opponents of gender equality and sexual democracy are seeking not only political but also epistemic power, producing a new body of gender knowledge' while at the same time delegitimising existing expertise and science. Following the January 2025 executive order 14151 on 'Ending Radical and Wasteful Government DEI Programs and Preferencing', which ended federal practices that considered DEI factors, goals, policies, mandates, or requirements, vocabulary linked to gender equality and women's rights have been deleted from US government websites, and US diplomats have pushed for deletion of the word 'gender' in UN resolutions.
A 23 July 2025 executive order 14319 on 'Preventing woke AI in the federal government' urged federal agencies not to procure services that 'sacrifice truthfulness and accuracy to ideological agendas'. The document included the explicit principle that LLMs 'do not manipulate responses in favor of ideological dogmas such as DEI'.
Men and boys in an AI‑accelerated information space
Gender plays a key role in targeted social media advertising and recommender systems which tailor different content to male and female users. According to the international men's health organisation Movember Foundation, two‑thirds of young men worldwide engage with 'masculinity influencers' online on a regular basis. According to a repplays a key role as a co-legislator in shaping all relevant regulation and overseeing the implementation and enforcement ort published by the Danish Centre for Digital Youth Care in October 2026, experiments with artificial profiles show that young men were exposed to misogynistic content in their feed after scrolling for just three minutes, on average.
As misogynistic narratives and incel ('involuntary celibacy') culture have moved from the fringe to the mainstream in recent years, many women are leaving male-dominated platforms such as X. Growing fragmentation of public debates – where individuals experience different social media environments depending on their gender – can have consequences in other ways: according to the Commission's 2025 strategic foresight report, an emerging political and ideological gap between young women and men is linked to deepening ideological echo chambers and polarisation on social media, where algorithms prioritising divisive content are pitting (especially young) men against girls and women. While direct causality has not been established, Flash Eurobarometer 544 on gender stereotypes and violence against women showed that, in nearly all questions, men aged 18‑24 and 25‑44 have more stereotyped perspectives and victim-blaming attitudes than older men.
At the same time, boys and young men are being targeted online by a growing industry of explicit 'AI girlfriends' and specialised 'nudifying' tools. They can upload photos of real girls and women and, within seconds, receive instant gratification by creating personalised non‑consensual AI sexualised material. A January 2026 report found that xAI's Grok had generated 3 million sexualised images, including 23 000 of children, within 11 days. Grok can then post the content on X, allowing users to monetise it, feeding into a sprawling economic ecosystem of online misogyny.
Parliamentary responses to NCII
The Inter‑Parliamentary Union (IPU) calls the challenge of AI‑generated NCII 'one of the clearest, most high‑profile tests of whether parliaments can shape the future of responsible AI'. In its June 2026 parliamentary action guide to combating NCII, the IPU sets five priorities for parliamentary action:
-
Legislate to recognise AI‑generated NCII as gender‑based violence.
-
Ensure effective takedown and victim support mechanisms.
-
Use oversight to track implementation, evidence and platform compliance with legal frameworks and international obligations.
-
Speak out publicly on NCII, building cross‑party support in parliament and including the issue in national action plans around technology‑facilitated gender‑based violence.
-
Cooperate internationally, including via the IPU, on a borderless harm whose perpetrators, victims and distribution mechanisms can sit in different jurisdictions.
The IPU specifically suggests that parliamentary members globally consider: drawing on international frameworks – including the Convention on the Elimination of All Forms of Discrimination against Women (CEDAW) – when working towards domestic regulation of NCII; accelerating the negotiation, signing and ratification of relevant instruments, including the United Nations Convention on Cybercrime, to close cross‑border enforcement gaps that cannot be addressed by domestic legislation alone.
According to a 2025 Common Sense Media report, 72 % of US teenagers aged 13‑17 use AI companions, which are calibrated for emotional interactions, creating cognitive and emotional bonds and gaining the trust of the users. According to a 2025 Center for Democracy & Technology survey, 19 % of US high‑school students had used AI chatbots for a romantic relationship. A 2024 European Tech Insights report showed that 26 % of adult European men and 15 % of adult women would consider having a romantic relationship with an AI companion.
Nudification apps – generative AI tools that can create sexualising deepfakes – use addictive design to keep users engaged, by triggering high doses of dopamine. When users share the content they have produced, new rewards are released when users receive social validation from peers. Constant high doses of dopamine can impact brain pathways, affect the prefrontal cortex, and cause addictions that can make users more vulnerable to other addictions. Adolescents are particularly vulnerable: research suggests that in response to the arousal and novelty of AI pornography, adolescent brains produce higher spikes of dopamine, making them chase new hits repeatedly. Once the neural pathway between AI and sexual pleasure has been established, real‑life sexual experiences can become less pleasurable in comparison. This can further contribute to the emotional and physical isolation of young men, making them even more vulnerable to incel narratives and monetised parasocial relationships. These vulnerabilities and compulsive behaviours can be exploited and monetised by actors that promote other addictive designs, such as gambling, betting and crypto scams.
The EU's response
The external dimension
The EU's political and operational roadmap towards a gender‑equal world is laid out in the EU gender equality strategy 2026-2030 and the EU gender action plan III 2021-2025 (extended until the end of 2027). Key actions by the Commission include a recommendation on safety in politics (2026); updating the mapping of legal and policy measures to promote the participation and leadership of women in politics, public administration and parliaments (2027); and a study on online networks, spheres and narratives targeting young men and boys (2028). The Commission will also organise a policy practice joint event on misogyny and radicalisation in November.
On a multilateral level, the EU and its Member States have partnered with the UN, and helped drive UN resolutions to counter violence against women and increase political and civic participation of women and girls. The EU actively implements UN Security Council Resolution 1325 on Women, Peace and Security (WPS Agenda) as a core principle of the common foreign and security policy. The Strategic Compass underlines the commitment to integrate gender equality and human rights in the common security and defence policy.
On 1 October 2023, the EU ratified the Council of Europe Convention on preventing and combating violence against women and domestic violence (Istanbul Convention), committing to adopting a gender perspective in all its measures to prevent violence against women, punish perpetrators and protect victims.
In its efforts to counter FIMI, including disinformation, the EEAS has also focused on identity‑based disinformation, including in its November 2024 guidelines on open source intelligence investigations on identity‑based disinformation/FIMI.
The democracy dimension
The 12 November 2025 communication on the European Democracy Shield (JOIN(2025) 791 final) included proposals to counter FIMI and disinformation. Against a backdrop of political candidates and elected representatives – in particular, women – being increasingly exposed to threats, attacks and violence, online and offline, with clear negative effects on democratic representation, the Commission announced a recommendation on safety in politics for political candidates and elected representatives concerning offline and online threats, including disinformation. The Commission also announced that it would work with the European cooperation network on elections (ECNE) on a guide to best practices, and would support capacity‑building, collection of data, and voluntary commitments as well as ethical standards. The European Digital Media Observatory (EDMO) has published several reports on disinformation against women and LGBTQ+ people in the EU. EDMO has also offered training on identity-based disinformation.
The digital dimension
Under the Digital Services Act (DSA), very large online platforms (VLOPs) – with over 45 million monthly users in the EU – are obliged to identify and analyse widespread, systemic risks. This includes protecting users against gender‑based violence and manipulated sexually explicit images, including child sexual abuse material (CSAM). Once identified, VLOPs must reduce the risks. VLOPs used for the dissemination to the public of pornographic content must ensure that victims can effectively exercise their rights to have content representing non‑consensual sharing of intimate or manipulated material removed without undue delay.
The Artificial Intelligence (AI) Act
Transparency obligations under the AI Act, fully applicable from August 2026, include mandatory labelling of AI‑generated content (i.e. watermarking). In June 2026, the Commission published a new Code of Practice on Transparency of AI-Generated Content, signed by 190 companies and organisations by the end of July 2026. The code supports marking and labelling of AI‑generated content in machine‑readable formats to enable detection. It also assists deployers using deepfakes or AI‑generated content to clearly disclose AI involvement.
In May 2026, the EU agreed to ban AI systems that generate non‑consensual sexually explicit and intimate content or child sexual abuse material, such as AI nudification apps. The ban will enter into force on 2 December 2026.
Directive on Combating Violence against Women
Article 5(1) of the EU Directive on Combating Violence against Women (2024/1385), adopted in 2024, mandates Member States to criminalise, by June 2027, key acts of gender‑based cyber violence such as cyber‑harassment, cyber incitement to violence and hatred, and the non‑consensual sharing of intimate images. The latter includes image‑based sexual abuse that can be facilitated by AI, such as deepfakes: a) 'making accessible to the public, by means of information and communication technologies ("ICT"), images, videos or similar material depicting sexually explicit activities or the intimate parts of a person, without that person's consent'; b) non‑consensual production, manipulation, altering, or subsequent dissemination of 'material making it appear as though a person is engaged in sexually explicit activities'; c) 'threatening to engage in the unlawful conduct referred to in point (a) or (b) in order to coerce a person to do, acquiesce to or refrain from a certain act'.
Role of the European Parliament
Parliament plays a key role as a co‑legislator in shaping all relevant regulation and overseeing the implementation and enforcement of EU regulation on AI, including through the Working Group on the Implementation of the Digital Services Act, and the Working Group on the Implementation and Enforcement of the AI Act.
In the findings and recommendations of the Special Committee on the European Democracy Shield (2025/2069(INI)) – adopted on 15 September 2026 by 420 votes to 220, with 26 abstentions – Members stressed that the safety and security of women in political life are prerequisites for equal representation and resilient democracy; and condemned disinformation campaigns, including via deepfakes and non‑consensual sexual content, used to harass, threaten and intimidate female candidates, frequently relying on narratives challenging their competence, credibility and moral integrity, thereby discouraging women's political participation and undermining trust in democratic institutions. They also called on the Commission to include in the forthcoming recommendations on safety in politics specific provisions on the monitoring, prevention and countering of disinformation campaigns specifically targeting female candidates.
During Parliament's sixth Gender Equality Week in November 2025, the Committee on Women's Rights and Gender Equality (FEMM) and EUDS organised a joint exchange of views on 'Radical movements advocating against gender equality such as manosphere, causing interference with European democratic processes'. In February 2026, Parliament adopted a resolution (2025/2240(INI)) with recommendations to the Council for the 70th UN Commission on the Status of Women (UNCSW) session. The resolution highlighted the need for a deeper understanding of anti‑rights and anti‑gender movements, including disinformation on gender equality, the incel phenomenon and misogyny, and particularly their online manifestation known as the 'manosphere'. Members also advocated more active policies to counter these phenomena and their narratives.
Parliament's 16 June 2026 resolution on countering transnational repression – towards an EU strategy to protect Europe's sovereignty and democratic values (2025/2179(INI)), called on 'the Commission and the Member States to adopt gender‑sensitive protection measures for women and LGBTIQ+ communities affected by TNR' and stressed 'the need to recognise the specific challenge of gender‑based digital TNR and the particular consequences of this type of repression for women targeted by TNR'.
Endnotes
Classification
Policy areas: Gender Issues, Equality and Diversity | Democracy
Regions: European Union
Committees: Legal Affairs (JURI), Constitutional Affairs (AFCO)
Acknowledgements
Statement on the use of AI
Any AI‑generated content in this text has been reviewed by the author.
Disclaimer
This document is prepared for, and addressed to, the Members and staff of the European Parliament as background material to assist them in their parliamentary work. The content of the document is the sole responsibility of its author(s) and any opinions expressed herein should not be taken to represent an official position of the Parliament.
