rtificial intelligence (AI) is rapidly reshaping how governments and institutions address international peace and security. AI systems are now used in border management, surveillance, humanitarian operations, conflict analysis, and information governance. Multilateral actors increasingly frame AI as a tool for faster, smarter, and more efficient responses to instability. The opportunities are significant, but so are the risks.
In July 2026, the United Nations will convene the Global Dialogue on AI Governance in Geneva, reflecting growing recognition that decisions about AI can no longer remain concentrated among major technology powers or private companies alone. The initiative seeks to create a more inclusive international platform where governments, civil society, academia, and the private sector can deliberate on the future of AI governance. But whether this effort succeeds will depend on whether it addresses power, not merely technology.
So far, global AI debates have focused on innovation, safety, interoperability, ethics, and competitiveness. These are important concerns, but the conversation remains overwhelmingly technocratic. AI governance is still largely framed as a matter of managing technological risk rather than addressing how power is distributed and exercised. That is the deeper governance problem.
Technology does not enter a neutral world. AI systems are deployed in societies already shaped by inequality, militarization, weak accountability, and uneven political power—conditions that are even more pronounced in conflict–affected settings. When institutions treat AI systems as neutral, they risk amplifying the very inequalities they claim to address.
This is where the Women, Peace, and Security (WPS) framework becomes essential not as a secondary inclusion mechanism but as a core governance foundation. The WPS agenda, established through UN Security Council Resolution 1325, is a global framework that centers women's participation, protection, and leadership in conflict prevention, peacebuilding, and security governance. Many policy discussions approach AI as a new issue that WPS institutions simply need to absorb. The relationship should be reversed. The international community does not merely need AI policies that consult WPS actors; it needs AI governance grounded in WPS principles from the outset.
Armed conflict is deeply gendered, but AI systems often flatten lived realities into classifications, probabilities, and predictive outputs. Women in conflict settings experience forms of insecurity that formal security institutions frequently fail to capture: retaliation for mediation work, harassment tied to online advocacy, coercion during displacement, reputational violence, and exclusion from public decision–making. These are not peripheral experiences. They are structural features of insecurity.
AI systems, however, operate through abstraction. They convert complex social realities into standardized categories derived from historical data and institutional assumptions. What cannot be easily measured often becomes invisible. This creates a dangerous governance gap.
For example, a predictive system assessing “security risks” may prioritize state–centric indicators while overlooking informal mediation networks led by women. Automated monitoring tools may flag activism or digital organizing as suspicious without understanding local political conditions. Humanitarian systems optimized for efficiency may unintentionally expose vulnerable communities to greater surveillance and visibility.
The issue is not simply algorithmic bias. The deeper problem is institutional overreliance on systems optimized for legibility and scale rather than for social understanding and care.
Used responsibly, AI can support peacebuilding efforts. Translation tools can reduce language barriers, while data aggregation and pattern–recognition systems can assist analysts in processing large volumes of information and contribute to early warning analysis under meaningful human oversight.
The danger emerges when institutions begin substituting political judgment with automated outputs. A system that supports human analysis is fundamentally different from one that classifies communities, generates threat assessments, automates surveillance, or shapes access to protection through opaque decision–making processes. Yet many institutions are adopting AI tools without clearly defining which functions must remain under accountable human authority. Political judgments are increasingly reframed as technical decisions. Institutions start to defer rather than decide.
For women peacebuilders, this shift has direct consequences. It shapes which warnings are taken seriously, which movements are monitored, which leadership becomes visible to surveillance systems, and which insecurity remains excluded because it does not fit standardized categories. This is precisely why WPS cannot remain a peripheral within AI governance debates.
The international community does not need AI systems retrofitted later to accommodate women’s participation. It needs governance frameworks fundamentally grounded in the principles established by the WPS agenda decades ago.
If institutions are serious about responsible AI in conflict settings, several guardrails must be explicit:
- Women must hold real technical and institutional decision–making authority in AI oversight, including monitoring, auditing, and evaluation roles. Advisory inclusion without authority will not alter outcomes.
- AI deployments in conflict settings must undergo gender–sensitive risk assessments before implementation, not after harm occurs.
- Functions affecting liberty, surveillance, or access to protection must never be fully automated. If institutions cannot explain how a system produces outputs, audit its assumptions, or provide meaningful avenues for contestation, that system should not shape security decisions.
- Multilateral organizations, peace operations, and humanitarian agencies must establish clear red lines of institutional responsibility when AI tools are used. Diffused accountability is incompatible with credible protection policy.
The emerging global AI governance architecture offers an opportunity to move beyond narrow conversations about inclusion and ethics toward deeper questions of governance itself. The issue is not whether women should have a seat at the AI table. The issue is whether AI governance frameworks will continue reproducing centralized and technocratic models that peacebuilding practitioners have spent decades trying to challenge.
WPS should not be treated as an add–on to AI governance. It should help define its foundation. Otherwise, institutions risk building systems that are faster, more scalable, and more efficient at reproducing the very inequalities they claim to solve.
a global affairs media network
Women, Peace, and Security is the missing core of AI governance
David Yu via Pexels.
May 27, 2026
AI does not enter a neutral world; when deployed in conflict–affected settings can amplify insecurities. Existing on–the–ground frameworks like Women, Peace, and Security that understand local inequalities and realities should act as a foundation for AI governance, writes Jamila–Aisha P. Sanguila.
rtificial intelligence (AI) is rapidly reshaping how governments and institutions address international peace and security. AI systems are now used in border management, surveillance, humanitarian operations, conflict analysis, and information governance. Multilateral actors increasingly frame AI as a tool for faster, smarter, and more efficient responses to instability. The opportunities are significant, but so are the risks.
In July 2026, the United Nations will convene the Global Dialogue on AI Governance in Geneva, reflecting growing recognition that decisions about AI can no longer remain concentrated among major technology powers or private companies alone. The initiative seeks to create a more inclusive international platform where governments, civil society, academia, and the private sector can deliberate on the future of AI governance. But whether this effort succeeds will depend on whether it addresses power, not merely technology.
So far, global AI debates have focused on innovation, safety, interoperability, ethics, and competitiveness. These are important concerns, but the conversation remains overwhelmingly technocratic. AI governance is still largely framed as a matter of managing technological risk rather than addressing how power is distributed and exercised. That is the deeper governance problem.
Technology does not enter a neutral world. AI systems are deployed in societies already shaped by inequality, militarization, weak accountability, and uneven political power—conditions that are even more pronounced in conflict–affected settings. When institutions treat AI systems as neutral, they risk amplifying the very inequalities they claim to address.
This is where the Women, Peace, and Security (WPS) framework becomes essential not as a secondary inclusion mechanism but as a core governance foundation. The WPS agenda, established through UN Security Council Resolution 1325, is a global framework that centers women's participation, protection, and leadership in conflict prevention, peacebuilding, and security governance. Many policy discussions approach AI as a new issue that WPS institutions simply need to absorb. The relationship should be reversed. The international community does not merely need AI policies that consult WPS actors; it needs AI governance grounded in WPS principles from the outset.
Armed conflict is deeply gendered, but AI systems often flatten lived realities into classifications, probabilities, and predictive outputs. Women in conflict settings experience forms of insecurity that formal security institutions frequently fail to capture: retaliation for mediation work, harassment tied to online advocacy, coercion during displacement, reputational violence, and exclusion from public decision–making. These are not peripheral experiences. They are structural features of insecurity.
AI systems, however, operate through abstraction. They convert complex social realities into standardized categories derived from historical data and institutional assumptions. What cannot be easily measured often becomes invisible. This creates a dangerous governance gap.
For example, a predictive system assessing “security risks” may prioritize state–centric indicators while overlooking informal mediation networks led by women. Automated monitoring tools may flag activism or digital organizing as suspicious without understanding local political conditions. Humanitarian systems optimized for efficiency may unintentionally expose vulnerable communities to greater surveillance and visibility.
The issue is not simply algorithmic bias. The deeper problem is institutional overreliance on systems optimized for legibility and scale rather than for social understanding and care.
Used responsibly, AI can support peacebuilding efforts. Translation tools can reduce language barriers, while data aggregation and pattern–recognition systems can assist analysts in processing large volumes of information and contribute to early warning analysis under meaningful human oversight.
The danger emerges when institutions begin substituting political judgment with automated outputs. A system that supports human analysis is fundamentally different from one that classifies communities, generates threat assessments, automates surveillance, or shapes access to protection through opaque decision–making processes. Yet many institutions are adopting AI tools without clearly defining which functions must remain under accountable human authority. Political judgments are increasingly reframed as technical decisions. Institutions start to defer rather than decide.
For women peacebuilders, this shift has direct consequences. It shapes which warnings are taken seriously, which movements are monitored, which leadership becomes visible to surveillance systems, and which insecurity remains excluded because it does not fit standardized categories. This is precisely why WPS cannot remain a peripheral within AI governance debates.
The international community does not need AI systems retrofitted later to accommodate women’s participation. It needs governance frameworks fundamentally grounded in the principles established by the WPS agenda decades ago.
If institutions are serious about responsible AI in conflict settings, several guardrails must be explicit:
- Women must hold real technical and institutional decision–making authority in AI oversight, including monitoring, auditing, and evaluation roles. Advisory inclusion without authority will not alter outcomes.
- AI deployments in conflict settings must undergo gender–sensitive risk assessments before implementation, not after harm occurs.
- Functions affecting liberty, surveillance, or access to protection must never be fully automated. If institutions cannot explain how a system produces outputs, audit its assumptions, or provide meaningful avenues for contestation, that system should not shape security decisions.
- Multilateral organizations, peace operations, and humanitarian agencies must establish clear red lines of institutional responsibility when AI tools are used. Diffused accountability is incompatible with credible protection policy.
The emerging global AI governance architecture offers an opportunity to move beyond narrow conversations about inclusion and ethics toward deeper questions of governance itself. The issue is not whether women should have a seat at the AI table. The issue is whether AI governance frameworks will continue reproducing centralized and technocratic models that peacebuilding practitioners have spent decades trying to challenge.
WPS should not be treated as an add–on to AI governance. It should help define its foundation. Otherwise, institutions risk building systems that are faster, more scalable, and more efficient at reproducing the very inequalities they claim to solve.
