UN Women

The success of the first Global Dialogue on AI Governance will depend on whether it delivers outcomes that are practical, measurable, and implementable. It must move beyond broad principles toward action that addresses the real and gendered impacts of AI. Online misogyny and TF VAWG are key drivers of and often a gateway to radicalization and violent extremism. In this sense, AI-enabled violence against women is not only a gender issue but a structural threat to inclusive governance, human rights, and social cohesion. At present, gender considerations remain largely absent from AI governance discussions, policy development, and across many stages of the AI lifecycle. The Global Dialogue offers an important opportunity to set a new standard by ensuring the meaningful participation of women and girls as experts, decision-makers, innovators, and rights-holders. Advancing gender-responsive AI governance means creating an inclusive, participatory process that actively engages women's rights movements, feminist organizations, civil society, academia, the private sector, and Member States to shape the agenda and outcomes. The Dialogue should catalyze multistakeholder commitments in the following areas: 1)Strengthen laws, regulatory frameworks, and accountability mechanisms that place a clear responsibility on technology companies and platforms to prevent, identify, and address negative impacts of AI systems on women and girls. Systematic use of gender-responsive impact assessments should be used across the AI lifecycle to identify, mitigate, and remedy differentiated risks and harms. 2) Advance safety by design approaches, including regulatory and policy measures, to prevent AI-enabled gender-based violence, harassment, exploitation, and discrimination. as evidence shows AI is accelerating images-based abuse (including deepfake abuse, impersonation, sextortion) coordinated harassment such as gendered disinformation and hate speech; doxing and violent threats data driven stalking,- all disproportionately affecting women and girls. 3)Strengthen investment in inclusive AI infrastructure, digital connectivity, and equitable access, particularly in low and middle income countries, so that AI reduces rather than widens existing divides. 4) Support capacity building that enhances ownership, participation, and leadership in AI governance, including technical skills and the meaningful engagement of women, youth, and marginalized groups. 5) Improve data governance, representativeness, transparency, and accountability to address structural biases embedded in AI systems. 6) Ensure women's leadership and multidisciplinary expertise are embedded in AI oversight, standards setting, monitoring, and enforcement at national, regional, and global levels, aligned with frameworks such as the Global Digital Compact and CSW67 Agreed Conclusions. The Dialogue should establish a pathway toward interoperable, credible, and human rights based international AI standards. While voluntary frameworks are useful, stronger mechanisms are essential for accountable global AI governance where risks are severe.

• Safe, secure and trustworthy AI
• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights

The priority areas constitute essential building blocks of responsible AI development and governance. They align with the UNESCO Recommendation on the Ethics of Artificial Intelligence and with feminist principles for gender equality in the Global Digital Compact, which emphasize human rights, inclusion, accountability, and the fair distribution of power and opportunity across digital ecosystems. They are also critical to ensuring that AI advances, rather than undermines, gender equality and the empowerment of women and girls. Together, these priorities address the impacts of AI related policy choices on people's lives. They provide guardrails to ensure AI systems are safe, transparent, accountable, and responsive to social realities, while preventing harm, discrimination, and the reinforcement of structural inequalities. They also require safeguards against the misuse of digital safety, cybersecurity, or public order laws by States to criminalize legitimate expression, expand surveillance, and monitor, silence, or intimidate women, including women human rights defenders and journalists . This is especially important for women and girls, who are disproportionately affected by online abuse, biased algorithms, digital exclusion, and underrepresentation in technology design and leadership. Safe, secure, and trustworthy AI is essential to prevent AI enabled violence, harassment, surveillance, and discriminatory outcomes. AI capacity building is equally important to equip countries and communities, especially women and girls, with the skills, resources, and leadership opportunities needed to shape AI futures rather than be excluded from them. Addressing the social, economic, ethical, cultural, linguistic, and technical implications of AI helps ensure that diverse realities are reflected in data, design, and governance. This is vital for women in all their diversity, particularly those facing intersecting forms of discrimination. Protection and promotion of human rights provides the foundation for privacy, dignity, equality, freedom of expression, and access to remedy. Importantly, these priorities recognize that AI impacts communities unevenly, shaped by access to resources, representation, language, geography, and historical exclusion. Embedding them into AI governance strengthens the potential of AI to contribute to sustainable and peaceful development which would benefit to all.

A key issue that is not sufficiently captured in the listed themes is a strong, dedicated focus on gender equality and specifically technology-facilitated violence against women and girls (TFVAWG) within AI governance frameworks. Aligned with UN Women's position paper to the Global Digital Compact, and consistent with the Feminist Principles for Including Gender in the Global Digital Compact (led by civil society and supported by UN Women and partners through the Gender in Digital Coalition), a dual track approach is essential: 1) a stronger, explicit focus on gender equality, 2) combined with robust mainstreaming across all AI governance clusters. Ensuring a gender lens across the full AI lifecycle is critical, including design, data collection, model development, deployment, procurement, and oversight. Without this, existing inequalities risk being embedded, scaled, and normalized through AI systems in ways that are often invisible within current governance approaches. This reflects deeper structural power asymmetries in the AI ecosystem, including who designs and controls technologies, whose data is used, whose knowledge systems are prioritized, whose ideas are funded and whose risks are recognized. Without explicit gender-responsive safeguards embedded across governance frameworks, AI risks reinforcing rather than correcting existing inequalities. Addressing this requires more than technical fixes. It demands the integration of gender-responsive requirements into standards, regulatory frameworks, auditing mechanisms, procurement systems, and accountability structures. It also requires meaningful participation and leadership of women and girls in all their diversity in all stages of AI governance, from agenda-setting and design to monitoring and remedy. Importantly, this must be treated both as a standalone priority and as a cross-cutting imperative across all AI governance pillars. Embedding a strong gender lens throughout the AI ecosystem is essential to ensure AI advances equality, strengthens accountability, and upholds human rights across all contexts and applications. Such an approach can also ensure that AI is better leveraged for positive social change including for the prevention of and response to violence against women and girls as well as fostering positive social norms.

At present, AI governance and development efforts continue to insufficiently recognize the disproportionate inequalities that artificial intelligence produces and amplifies in the lives of women and girls. AI systems and governance frameworks are currently reproducing and scaling structural gender inequality through exclusion in design, gender‑blind policy, biased data, and weak accountability mechanisms. These dynamics contribute to a widening gendered AI divide, driven by connectivity gaps, unequal access to digital skills, and extractive data practices, among other exclusionary technology architectures that further constrain women's agency and leadership in an AI‑driven future. When gender considerations are absent from system design and public policy, existing power asymmetries are reproduced and amplified at scale. This is particularly consequential as AI systems are increasingly embedded in pathways shaping access to employment, finance, public services, justice and civic participation. Two critical gaps shape these outcomes across the AI lifecycle. First, women remain significantly underrepresented in AI‑related education and employment, accounting for approximately 30 per cent of AI professionals. This underrepresentation influences which problems are prioritized, which risks are considered material, and whose experiences are reflected in AI systems. Second, gender responsiveness in national AI strategies remains limited, with only a small minority of countries referencing gender considerations and even fewer including substantive gender‑responsive provisions. Exclusion in AI design increases the likelihood that systems will reproduce gender stereotypes and disadvantage women across all aspects of AI interactions, with compounded impacts for those facing intersecting forms of discrimination. At the same time, AI presents significant opportunity when we meaningfully close the gender divide. Closing this divide could generate substantial economic gains and lift 30 million women and girls out of poverty by 2050. UN Women calls for AI governance and development that empowers all, not just selected few.

The AI Dialogue plays an important role in advancing international cooperation on artificial intelligence governance by providing a central, Member State–led platform for inclusive, transparent, and constructive engagement. It serves as a forum through which governments, International Organisations, the private sector, academia, civil society, and other relevant stakeholders can exchange views, share expertise, and build mutual understanding in a spirit of cooperation and multilateralism. The credibility and legitimacy of the AI Dialogue are rooted in its grounding within a Member State–driven process, consistent with the principles of the United Nations Charter, international law, and agreed international norms. The AI Dialogue contributes to fostering trust and facilitating convergence of perspectives on complex and rapidly evolving issues related to the development, deployment, and governance of AI technologies. Inclusive participation enables the Dialogue to reflect the diversity of national contexts, capacities, and development priorities, including those of developing countries and countries in special situations, thereby supporting efforts to narrow digital divides and promote equitable and sustainable outcomes. Ensuring the full, equal, and meaningful participation of women, in all their diversity, must be a is a key component of the AI Dialogue and is essential to its effectiveness. Diverse perspectives strengthen the quality, legitimacy, and relevance of deliberations and outcomes, and contribute to AI governance approaches that are responsive to the needs of all societies. Importantly, the AI Dialogue can also play a norm-setting role by establishing a global benchmark for the meaningful inclusion of women and girls and the systematic integration of a gender lens across all AI governance discussions and outcomes. To serve its purpose as a primary forum for discourse on AI governance, the AI Dialogue must remain inclusive, democratic, and accountable, with transparent processes and meaningful opportunities for participation. Upholding these principles is essential for advancing coherent, legitimate, and effective international cooperation on AI governance.

The AI Dialogue should build upon and support implementation of the Global Digital Compact, while drawing on the Agreed Conclusions of CSW67, which recognized innovation and technological change as key enablers of gender equality and the empowerment of all women and girls. It should also connect with existing multi-stakeholder initiatives that have already generated practical commitments and partnerships, including the Generation Equality Action Coalition on Technology and Innovation for Gender Equality, a multi-stakeholder alliance of Governments, private sector actors, civil society, youth leaders, and International Organisations working to advance feminist technology policy and close gender gaps in innovation; the Global Partnership for Action Against Online Harassment and Abusee; and the Gender and Digital Coalition,. The Dialogue can add value by bringing these strands of work into a coherent global governance space focused specifically on artificial intelligence. It should serve as a forum to translate existing gender equality, human rights, and digital inclusion commitments into concrete AI governance measures across the full lifecycle of AI systems, from design and development to deployment, monitoring, and redress. This includes advancing shared commitments to ensure AI systems are transparent, reliable, safe, and subject to accountable human oversight, with fairness and accountability at their core. Consistent with these frameworks, the Dialogue should promote approaches that recognize the primary responsibility of Governments to identify, prevent, and address risks associated with AI systems, including discrimination, bias, exclusion, surveillance, and harm, while also recognizing the responsibilities of researchers, developers, and companies to assess, communicate, and mitigate such risks. The Dialogue should further support agile and coherent governance frameworks that combine international norms and guidance, national regulatory approaches, and technical standards, while enabling the exchange of lessons learned and emerging practices across sectors and regions. Finally, it should help ensure the full, equal, effective, and meaningful participation and leadership of women and girls in AI governance, and accelerate action to close the gender digital divide, promote digital safety, and expand women's participation in technology leadership and policymaking .

The AI Dialogue should provide meaningful and inclusive opportunities for participation across stakeholders, in line with the Global Digital Compact and the agreed conclusions of CSW67. Diverse constituencies must be visibly represented in both the process and outcomes of the AI Dialogue, particularly those most affected by the asymmetrical power dynamics of AI development, including women and girls, women's rights groups and digital rights activists and other marginalized groups. Inclusion must extend beyond presence to influence. The Dialogue should adopt an inclusive, multidisciplinary understanding of expertise that values lived experience, community knowledge, and human rights perspectives alongside technical expertise, moving away from narrow, elite, or technocratic definitions of who qualifies as an "expert." Consistent with CSW67's emphasis on inclusive decision‑making and the GDC's commitment to multistakeholder engagement, the Dialogue presents an opportunity to shift from consultative participation toward a more community‑driven AI discourse. Community‑driven approaches are essential to addressing the structural power of asymmetries and exclusionary architectures embedded in AI systems. When communities are meaningfully engaged across the AI lifecycle, including decisions on whether, how, and under what conditions data, languages, and knowledge are used for AI development, as well as having meaningful rights to refuse extractive practices, AI systems and governance frameworks are more likely to uphold human rights, advance gender equality, and distribute the benefits of digital transformation more equitably. Such engagement strengthens accountability, enhances legitimacy, and supports the transformative participation envisioned in both the Global Digital Compact, CSW 67 and the Beijing Declaration and Platform for Action.

The voices of marginalized constituencies, including women and girls, remain significantly underrepresented in AI governance and decision‑making processes. This exclusion reflects the exclusionary nature of prevailing AI architectures and infrastructure, as well as persistent structural gaps in AI education, employment, and policy. Women comprise only approximately 30 per cent of the AI workforce, shaping which problems are prioritized, which risks are treated as material, and whose interests are reflected in AI systems. These inequalities are reinforced at the policy level. Gender responsiveness in national AI strategies remains limited: among 138 countries assessed, only 24 reference gender considerations and only 18 include substantive gender‑responsive provisions. This gap undermines the Global Digital Compact's commitment to inclusive digital transformation and CSW67's call for gender‑responsive innovation and technological change. AI governance must therefore go beyond consultation and actively embed women's participation as decision‑makers and leaders across governance structures, standard‑setting bodies, and oversight mechanisms. Without women's leadership in shaping rules, priorities, and accountability frameworks, AI governance risks perpetuating the same power asymmetries it seeks to address. In AI development practices, access to critical inputs, such as computing resources, data, and access to AI tokens, remains concentrated among already privileged actors with access to infrastructure and technical capacity. Women and girls, who continue to face systemic barriers in connectivity, skills, and access to ICT infrastructure, are therefore further excluded from contributing to and shaping AI development. This compounded exclusion entrenches unequal power relations within AI systems and governance frameworks, reinforcing whose knowledge is valued and whose realities remain invisible.

The AI Dialogue should consider establishing regional and local participation hubs, including at United Nations premises, civil society partner sites, universities, and community centers, in advance of the main Dialogue dates. These hubs would enable stakeholders who are unable to travel in person, or who face connectivity or resource constraints, to participate meaningfully in the Dialogue process through facilitated discussions and collective deliberation. During the main Dialogue sessions, regional and local hubs should have the opportunity to present their key findings, priorities, and perspectives and to contribute through scheduled online interventions, ensuring that locally grounded inputs inform global discussions and outcomes. To further strengthen inclusive and community‑driven engagement, the AI Dialogue should consider allocating reserved agenda space in which topics and priorities are proposed by communities themselves, rather than exclusively by institutions or co‑facilitators. This approach would help anchor the Dialogue in lived experience and local knowledge, and support more bottom‑up, participatory AI governance processes. In addition, the AI Dialogue week could benefit from designating specific thematic days to address cross‑cutting priority issues. In this regard, consideration could be given to convening a dedicated Gender Day, drawing on the successful precedent of Gender Days in other multilateral processes, such as the Conference of the Parties under the UN Frameworks Convention on Climate Change. A dedicated Gender Day would provide focused space to advance gender‑responsive AI governance, highlight women's leadership and participation, and address gender‑based risks and impacts associated with AI systems.

UN Women advocates for AI governance centered on human rights and gender equality, as highlighted in the Global Digital Compact and CSW 67, which calls for gender mainstreaming and international cooperation to eliminate TFVAWG. Key policy and practice examples from the resource include: Normative Frameworks: General Assembly Resolution A/RES/79/152 urges Member States to refrain from using AI systems that are incompatible with international human rights law or pose undue risks to the enjoyment of those rights. It also calls for full, equal and meaningful participation of women in AI development and data management etc and preventing, addressing, prohibiting discrimination, intimidation, harassment and violence offline and online including in the data algorithms used in AI. Additionally, the CSW67 Agreed Conclusions and the UN Women Position Paper for the Global Digital Compact provide recommendations for governments to address technological change through a gender-responsive lens. Multistakeholder Collaboration: The Generation Equality Action Coalition on Technology and Innovation fosters catalytic action to bridge digital gender gaps and enhance online safety for women and girls. Data-Driven Solutions: Utilizing big data analytics provides concrete insights into online violence. For instance, a study in Libya in 2023 used these tools to provide social media authorities with evidence of where and how violence is proliferating, enabling targeted interventions. Similar big data monitoring has been used to track online misogyny and hate speech in Asia and the Pacific. Knowledge Standardization: Developing a Shared Research Agenda and common definitions for TFVAWG ensures that data collection is reliable and comparable, which is essential for informing effective AI safety policies and programmes. Safety by Design: UN Women promotes a "Safety by Design" approach to ensure that digital spaces and technologies are intentionally built to be safe and accessible for women, preventing the exacerbation of existing inequalities. Prevention and social norms transformation: Addressing the negative impacts of AI also requires upstream prevention strategies that challenge the gender norms and power dynamics driving online misogyny and abuse. UN Women works with men and boys as allies to transform harmful masculinities, including through research to better understand pathways into misogynistic and extremist networks online, the role of AI-driven amplification of harmful narratives, and entry points for disrupting these dynamics and fostering positive, gender-equal attitudes toward women and girls. These initiatives collectively work to close implementation gaps in regulation, data, and response mechanisms.