Artificial intelligence (AI) is often presented as the future of innovation. Every day, new tools are introduced that promise to transform how we work, communicate, and create. However, behind the excitement surrounding these technologies lies a more uncomfortable reality: AI does not exist separately from society. It develops within systems that already contain inequalities, discrimination, and uneven power relations. One of the clearest examples of this is the growing use of AI to spread gender-based disinformation.
Women politicians, journalists, activists, and human rights defenders have historically been attacked through narratives questioning their intelligence, morality, appearance, sexuality, and ability to lead.
This pattern is evident beyond high-profile international cases. A 2026 monitoring report on political misogyny in Armenia documented how women politicians were systematically targeted through what researchers describe as “gender punishment.” Rather than debating their political positions, online narratives focused on their motherhood, morality, sexuality, and personal lives. Women were portrayed as “bad mothers,” immoral, or unfit for public office, demonstrating how gendered attacks are designed to undermine political legitimacy rather than engage with policy.
What has changed is the technology behind these attacks. AI has made it easier to create convincing false information, manipulate images and videos, imitate voices, and produce large amounts of targeted content within minutes. Narratives that once required significant resources can now be created and distributed at an unprecedented speed. AI did not create misogyny. Women were targeted long before artificial intelligence existed. However, AI has amplified the systems that allow misogyny to spread by making these attacks faster, cheaper, and harder to challenge.
For example, during Armenia’s 2026 pre-election period, Women’s Resource Center Armenia documented coordinated misogynistic narratives targeting women politicians through sexist insults, sexualized language, and attacks on their roles as mothers. These campaigns existed independently of AI, but they demonstrate exactly the kinds of harmful narratives that generative AI can now reproduce, personalize, and distribute at a much greater scale.
The question we should therefore ask is not only what AI can create, but also who benefits from these technologies, who they harm, and who is held responsible when they are used as tools of intimidation.
What is gender-based disinformation?
Gender-based disinformation refers to false or misleading information that targets individuals because of their gender and uses sexist stereotypes to undermine their credibility. Unlike general disinformation, which may focus on political events or policies, gender-based disinformation attacks identity.
Women are often targeted through narratives about their identity and attributes. A woman politician may be attacked through rumours about her personal life instead of discussions about her political ideas. A woman journalist may be portrayed as emotional or dishonest rather than having her reporting questioned. Activists and human rights defenders are often labelled as dangerous, immoral, or influenced by foreign actors in an attempt to dismiss their work.
In their research about gendered disinformation as violence, Gehrke and Amit-Danhi (2025) have argued that gendered disinformation should not be understood only as a communication problem. It functions as a form of gender-based violence because it is used to intimidate, silence, and discourage women from participating in public life. These attacks are rarely only about one individual woman. They send a wider message: women who challenge traditional power structures will face consequences.
AI has changed the scale of gendered attacks
One of the clearest examples of AI’s amplification of gendered violence is the rise of deepfakes. AI-generated images and videos can make it appear as if someone said or did something that never happened. While deepfakes can target anyone, women are disproportionately affected by sexualized deepfake abuse.
Research shows that the majority of deepfake pornography targets women, often without their consent (see Saner, 2026; Alacantra & Valentim, 2023). These images are used to humiliate, threaten, and control victims. Once sexualized content spreads online, victims lose control over where it appears and who can access it.
In 2024, explicit AI-generated images of Taylor Swift spread widely online, gaining millions of views before being removed. The incident received global attention because of Swift’s public profile, but it also highlighted a larger issue: if one of the world’s most recognizable women cannot easily control the spread of AI-generated sexual abuse, what happens to women without the same visibility, resources, or legal support?
The problem is not only deepfakes. AI can also generate fake accounts, create large amounts of misleading content, and personalize attacks toward specific individuals. Combined with social media algorithms that reward engagement and controversy, harmful narratives can spread faster than corrections. The issue is therefore the unequal power between those who create harmful narratives and those forced to defend themselves against them.
Women are targeted because their voices challenge power
Gender-based disinformation often becomes most visible when women enter spaces traditionally dominated by men. During Germany’s 2021 federal election campaign, Annalena Baerbock, the only female candidate among the major parties, faced significant gendered attacks. Many narratives focused on her personality, appearance, and private life rather than her political positions. Instead of debating her ideas, these attacks questioned whether she was suitable for leadership.
A similar pattern appeared during Kamala Harris’s 2020 US vice-presidential campaign. Online narratives frequently relied on both sexist and racist stereotypes, portraying her as angry, unqualified, or successful because of personal relationships rather than professional achievements.
Women journalists also experience widespread online violence. UNESCO has documented that women journalists around the world face threats, harassment, and coordinated attacks, particularly when reporting on politics, human rights, or gender equality. The purpose of these campaigns is often not only to spread false information. It is to make participation exhausting.
A woman politician portrayed as immoral does not need to have her policies debated. A woman journalist portrayed as dishonest does not need to have her reporting considered. A woman activist portrayed as dangerous does not need to have her demands taken seriously.
Gender-based disinformation controls who is heard and who feels safe speaking.
AI does not remove inequality; it can automate it
One of the most concerning assumptions about AI is that technology itself is neutral.
AI systems are created within societies that already contain sexism, racism, and other forms of inequality and discrimination. When these systems are trained on existing data, they can reproduce and amplify the biases embedded within those social structures. This means AI is not simply a neutral tool that creates content; it can become a mechanism that strengthens existing power relations by enabling the rapid production and spread of narratives that reinforce gendered stereotypes and inequalities. In particular, AI-generated content can make harmful narratives appear more credible, widespread, and difficult to challenge, disproportionately affecting women and other marginalized groups.
Social media platforms further contribute to this problem. Many platforms are designed around engagement, meaning that content generating strong emotional reactions is often prioritized and given greater visibility. Misogynistic narratives can spread quickly because they are designed to provoke anger, fear, and outrage. A fabricated image, false accusation, or manipulated video can reach thousands of people before accurate information has the opportunity to challenge it.
The person targeted is left proving that something did not happen, while those responsible for creating the harm often face limited consequences or stay completely anonymous.
The democratic consequences of silencing women
Gender-based disinformation is a democratic issue. When women withdraw from public discussions because of harassment, threats, or fear of reputational damage, society loses important voices and perspectives.
Research shows that women targeted by online attacks often experience anxiety, fear, emotional exhaustion, and self-censorship. Some reduce their online presence, avoid certain topics, or leave public-facing roles entirely (Saner, 2026; Gehrke, 2025). This is exactly what makes gender-based disinformation dangerous and overbearing. Its success is measured by whether people believe a lie. It is also measured by whether a woman becomes tired enough, afraid enough, or isolated enough to stop participating.
Digital spaces have become central to politics, activism, journalism, and social movements. Being pushed out of these spaces has real-world consequences.
Building resistance: A feminist approach to AI
Addressing AI-driven gender-based disinformation requires proactive strategies that prevent harm, support those targeted, and challenge the systems that allow abuse to spread. Technology companies must take greater responsibility by developing stronger detection and moderation systems that can identify AI-generated harassment, deepfakes, and coordinated disinformation campaigns targeting women. They should also increase transparency around how algorithms amplify harmful content and create accessible reporting systems that provide meaningful support for individuals experiencing digital abuse.
Governments and institutions also have an important role in creating human rights-based policies that address technology-facilitated gender violence while protecting freedom of expression. Policies concerning women’s rights and artificial intelligence should not be developed separately but considered together, as AI systems can reproduce and amplify existing gender inequalities. A gender-sensitive approach to AI governance is necessary to address how women and marginalized groups are disproportionately affected by online harassment, algorithmic bias, and AI-generated abuse. By integrating gender perspectives into AI policies, societies can ensure that technological development does not reinforce existing power structures but instead contributes to safer and more inclusive digital spaces.
Protection and resistance also require empowering individuals and communities with digital literacy and practical tools to recognize, report, and respond to disinformation. Journalists, activists, and women in public life should have access to safety training, support networks, and resources that allow them to continue participating online without being forced to withdraw because of harassment or threats. However, the responsibility to create safer digital environments should not fall only on those experiencing harm. Women should not have to reduce their voices, limit their participation, or avoid public spaces because platforms and institutions fail to address abuse.
A feminist approach to AI requires accountability, community support, and a commitment to challenging the inequalities that allow gender-based disinformation to succeed. It means ensuring that women’s experiences are considered in the design, regulation, and implementation of AI systems. Through responsible technology development, inclusive policies, digital education, and collective action, societies can work toward digital spaces where women can participate freely, safely, and equally.
By Varanta Aharonian Niziblian
Further Reading
Alcantara, J., & Valentim, J. (2023). Gender-Based Disinformation: A scoping review of the literature, 2013–2023.
Gehrke, M., & Amit-Danhi, E. R. (2025). Gendered disinformation as violence: A new analytical agenda. Harvard Kennedy School Misinformation Review.
Gehrke, M. (2025). How are gender and race deployed in disinformation? An exploratory study of women politicians.
Saner, K. E. (2026). Introduction to the special issue on gendered disinformation.
