Janes

A successful first Global Dialogue on AI Governance should move beyond high-level principles to deliver actionable, interoperable pathways that connect AI governance to real-world security risks and human outcomes. In particular, it must recognise that AI is not only a technical system, but an information infrastructure shaping how conflicts are understood, prioritised, and acted upon. First, the Dialogue should explicitly address "narrative erasure" as a security vulnerability. Algorithmic amplification can distort visibility by silencing certain harms while sensationalising others; thereby influencing early warning, humanitarian prioritisation, and policy response. A meaningful outcome would be commitments from states and the private sector to identify and mitigate gendered and contextual biases in AI systems, strengthening information integrity as a pillar of safe and trustworthy AI. Second, success would require integrating an "Algorithmic Responsibility to Protect" framework, bridging technical AI governance with international law. This means recognising that AI-amplified, demographically targeted disinformation can trigger real-world violence, and that Member States have a responsibility to mitigate such risks. AI governance must therefore extend beyond "trustworthy systems" to include the protection of human life in AI-mediated conflict environments. Third, success would involve developing interoperable and inclusive governance mechanisms, particularly for the Global South. This includes building capacity not only for access to AI, but for OSINT integrity, enabling actors in conflict-affected regions to verify AI-generated content and reduce analytical blind spots. Ultimately, the Dialogue will be successful if it produces not just consensus, but implementable frameworks that align AI governance with conflict prevention, human dignity, and equitable global participation.

• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches
• Protection and promotion of human rights
• Safe, secure and trustworthy AI

My selection of these four priorities is driven by the urgent need to address the structural governance failures created by AI-amplified disinformation in conflict zones. As an APAC Manager in the defense intelligence sector and a researcher in AI security ethics, my focus is on how these thematic areas intersect to prevent kinetic violence. 1. Safe, Secure, and Trustworthy AI: For AI to be truly "safe," it must be resilient against the weaponization of gendered disinformation. My research into "Algorithmic Responsibility to Protect" argues that trustworthiness is not just a technical metric but a security requirement; AI systems must be designed to mitigate narrative manipulation that erases or weaponizes the identities of vulnerable groups. 2. Implications of AI (Social, Ethical, and Cultural): We must prioritize understanding the societal and ethical implications of AI-mediated narratives. In my work, I have seen how algorithmic bias can trigger real-world conflict by amplifying toxic tropes. Addressing these cultural and ethical dimensions is essential for conflict prevention and ensuring that AI deployment does not exacerbate existing regional tensions. 3. Interoperability of Governance Approaches: Current national digital regulations are often structurally unprepared for the borderless nature of AI-driven threats. We need interoperable governance that aligns international law with technical standards. This ensures that "Responsibility to Protect" frameworks can be operationalized across different jurisdictions to address security governance failures. 4. Protection and Promotion of Human Rights: AI governance must be rooted in the protection of human rights, specifically the right to safety from targeted digital violence. By centering human oversight and transparency, we can ensure that AI-driven Open Source Intelligence (OSINT) remains an instrument for truth and peace rather than a tool for narrative erasure and kinetic escalation.

While the listed themes are comprehensive, there is a critical emerging gap regarding "Asymmetric Narrative Warfare" and the "Securitization of Algorithmic Bias." Based on my research into Algorithmic R2P, I believe the following cross-cutting issues require urgent inclusion: The Governance of "Narrative Erasure": While the themes cover "linguistic implications," they miss the strategic risk of narrative erasure. In conflict zones, algorithms can be manipulated to systematically suppress the voices of vulnerable groups, creating a "governance failure" that current national digital regulations are structurally unprepared to address. Algorithmic Responsibility to Protect (R2P): We must move beyond "safe and secure" systems toward a framework of state and corporate accountability for atrocity prevention in the digital sphere. This cross-cuts security and ethics by defining the "duty to protect" populations from AI-mediated narrative weaponization. OSINT Integrity in Conflict Prevention: The dialogue focuses on "open-source software," but the emerging priority is Open-Source Intelligence (OSINT) Integrity. As an APAC Manager at Janes, I see that the reliability of intelligence used for conflict monitoring is under threat from AI-generated "hallucinations" and deepfakes. Ensuring the integrity of the data used for global security decision-making is a technical and diplomatic necessity not fully captured by "transparency" alone. Gendered Disinformation as a Kinetic Trigger: Existing themes treat disinformation as a "human rights" or "trust" issue. However, my work demonstrates that AI-amplified gendered disinformation acts as a mechanism of kinetic conflict, weaponizing demographic narratives to trigger physical violence and institutional collapse. By addressing these gaps, the Dialogue can transition from a tech-centric focus to a conflict-prevention framework that protects the most vulnerable in AI-mediated environments.

In the defense intelligence sector and across the Asia-Pacific, governance gaps are increasingly visible in how AI-mediated information ecosystems translate into real-world security risks. These challenges are particularly evident in conflict-prone and information-fragile environments. Key Challenges 1. Weaponization of Identity Narratives In contexts such as Myanmar and parts of Northeast India, gendered and identity-based disinformation is amplified by algorithmic systems, intensifying communal tensions. This reflects a governance gap where existing digital and security frameworks are not equipped to address the security implications of AI-driven narrative manipulation. 2. Intelligence Blind Spots through Narrative Erasure AI systems often underrepresent or mischaracterize the experiences of marginalized groups in conflict zones. This phenomenon, sometimes described as "narrative erasure", creates gaps in OSINT-based analysis, leading to incomplete situational awareness and weakened early warning capabilities. 3. Information Integrity and Synthetic Media Risks The rise of AI-generated content, including deepfakes, poses challenges to data verification and credibility, particularly in politically sensitive contexts such as elections. The absence of interoperable standards increases the risk of misinterpretation or escalation based on manipulated information. Opportunities for Governance and Practice These challenges also present opportunities to advance integrated and human-centered AI governance. Frameworks such as "Algorithmic R2P" can help link AI governance with conflict prevention by recognising disinformation as a security risk. Additionally, strengthening public-private collaboration, enhancing human oversight, and developing standards for trustworthy AI can improve analytical resilience and support more accurate, context-sensitive security assessments.

The AI Dialogue can serve as a critical multilateral platform to move from fragmented national approaches toward coherent and interoperable international governance frameworks. From a defense intelligence and Asia-Pacific (APAC) perspective, its role is particularly important in three areas. First, operationalising responsibility in AI-mediated conflict environments. In contexts such as Myanmar and the Rohingya crisis, social media–driven disinformation has been widely documented as contributing to communal violence. As AI systems increasingly shape information flows, the Dialogue can help establish norms that recognise AI-amplified disinformation as a security risk, clarifying state and corporate responsibilities in mitigating harm, particularly where digital narratives translate into real-world violence. Second, strengthening "intelligence integrity" in AI-enabled systems. In politically sensitive environments such as Bangladesh's electoral cycles or heightened tensions in the China–Taiwan information domain, the proliferation of AI-generated content and deepfakes creates risks of misinterpretation and escalation. The Dialogue can support the development of shared verification standards and risk assessment frameworks, helping analysts and institutions distinguish credible signals from manipulated content. Third, addressing narrative bias and visibility gaps. Across parts of Oceania and other underrepresented regions, limited data representation can result in AI systems overlooking local vulnerabilities, including climate-linked security risks or gendered impacts of crises. The Dialogue can foster cooperation between Member States and the private sector to ensure AI systems are more inclusive, context-sensitive, and representative, reducing "narrative erasure" and improving situational awareness. Ultimately, the AI Dialogue can shift international cooperation from reactive coordination toward preventive, context-aware governance, aligning AI systems with conflict prevention, human rights, and regional stability.

To ensure the Global Dialogue on AI Governance succeeds, it must avoid duplication by connecting with established security and ethics frameworks. Based on my research into Algorithmic R2P and regional security in the Asia-Pacific, the Dialogue should build upon the following mechanisms: First, the Dialogue should connect with UNIDIR's AI and Security Ethics (AISE) programme, which has already convened governments, industry actors (including engagement with partners such as Microsoft), and researchers on the security implications of AI. The Dialogue can scale these discussions from technical and policy exchanges into broader multilateral norms, particularly in areas such as AI use in defence and conflict settings. Second, the Global Partnership on AI (GPAI) provides important technical expertise and working groups on responsible AI. However, its outputs are not always fully integrated into multilateral decision-making. The Dialogue can act as a bridge between GPAI's technical recommendations and the wider UN membership, ensuring that governance approaches are interoperable and inclusive, particularly for developing countries. Third, existing UN security frameworks such as the Programme of Action on Small Arms and Light Weapons (UN PoA) and regional mechanisms like UNRCPD offer entry points for linking AI governance with conflict prevention. While not designed for AI, these frameworks highlight how information dynamics and technology can influence escalation pathways, including the spread of violence. The added value of the Dialogue lies in its ability to connect these otherwise siloed efforts. It can advance practical cooperation by: - Integrating AI governance with conflict prevention and security frameworks - Supporting shared approaches to information integrity and risk assessment - Translating high-level principles into context-sensitive, operational guidance In doing so, the Dialogue can act as a coordinating platform, aligning technical, policy, and security communities toward more coherent and preventive AI governance.

The AI Dialogue must move beyond high-level diplomatic statements toward a practitioner-led structure that incorporates technical, security, and ethical perspectives. As a Manager in the defense intelligence sector, I recommend a Tri-Sectoral Working Group format: a. Multilateral & State Actors: Focus on standardizing Interoperable Governance Approaches to align national digital regulations with international law, specifically concerning the Responsibility to Protect (R2P). b. The Private Sector (Tech & Intel): Organizations like Janes and Microsoft should lead technical tracks on OSINT Integrity. Their role is to establish standards for verifying AI-generated content to prevent kinetic escalation in conflict zones. c. Civil Society & Academia: These stakeholders must act as "Ethical Auditors," identifying algorithmic biases and narrative erasures that state-centric models might overlook. From a defense intelligence perspective, practitioners working with OSINT and conflict analysis should be included more systematically, as they engage directly with AI-mediated information environments and can highlight real-world risks such as narrative distortion and early warning failures. In terms of structure, the Dialogue would benefit from a three-tiered format: - Thematic plenaries to align on principles and priorities - Focused working groups (e.g., AI and conflict, information integrity, human rights) to develop actionable recommendations - Implementation roundtables to translate discussions into policy tools, standards, or pilot initiatives Embedding feedback loops between these layers is essential to ensure continuity.

The most significant gap in current AI governance is the "Global South Practitioner" perspective, specifically those at the intersection of security and gender in the Asia-Pacific. Currently underrepresented voices include: a. Regional Security Analysts from the Global South: As a researcher from India, I observe that global standards often lack the hyper-local context of how AI-mediated narratives weaponize demographic tropes in regions like Myanmar or Northeast India. b. Gender & Conflict Specialists: The security implications of gendered disinformation are often relegated to "social issues" rather than being treated as the kinetic triggers they are. c. Grassroots Verification Networks: Those on the front lines of countering AI-generated deepfakes in developing nations lack a seat at the high-level policy table. To address this, inclusion must move beyond representation toward structural participation: The Dialogue should implement a "Regional Hub" model, hosting satellite consultations in cities like Kathmandu or Bengaluru. Furthermore, providing fully funded sponsorship for selected researchers from the Global South is essential to ensure that "inclusive participation" is not a luxury afforded only to the Global North. - Dedicated consultation tracks for conflict-affected and Global South stakeholders - Partnerships with regional organisations and local civil society networks - Funding and capacity support to enable meaningful participation - Mechanisms to integrate practitioner insights into policy discussions

To foster dynamic engagement, the Dialogue should adopt formats that simulate the complex, real-world impacts of AI governance failures: Algorithmic Red-Teaming Simulations: Instead of static panels, stakeholders should participate in a Conflict-Simulation Lab. This involves using hypothetical scenarios where AI-amplified disinformation triggers a regional security crisis. This format allows diplomats and tech leads to test the efficacy of "Algorithmic R2P" frameworks in real-time. Interactive Narrative Exhibitions: The Dialogue should feature interactive digital galleries that visualize Narrative Erasure. Showing how specific groups are systematically erased from algorithmic models makes the abstract concept of bias a tangible security concern for policymakers. The Reverse Mentorship Roundtable: Pair senior UN diplomats with early-career tech practitioners and analysts to bridge the tech-policy literacy gap. Hybrid Lightning Sprints: Focused 60-minute sessions dedicated to solving a specific governance gap (e.g., "Standardizing Deepfake Verification in OSINT"). In addition, regional insight roundtables would allow participants from specific contexts (e.g., Asia-Pacific, Africa) to present grounded experiences, ensuring that global discussions are informed by local realities. These formats move the Dialogue away from reporting on AI risks and toward actively architecting a secure and ethical future.

Several existing policies, standards, and practices offer concrete pathways for effective AI governance, particularly when they combine technical tools, policy frameworks, and operational safeguards. At the technical level, the Coalition for Content Provenance and Authenticity (C2PA), supported by industry actors including Microsoft, provides a practical approach to strengthening information integrity. By embedding verifiable metadata into digital content, it enables traceability and helps address the growing challenge of AI-generated synthetic media, particularly in conflict-prone environments. At the policy level, frameworks such as the UNESCO Recommendation on the Ethics of AI and the OECD AI Principles establish widely endorsed norms around human rights, transparency, and accountability. The EU AI Act further operationalises these principles through a risk-based regulatory model, offering a structured approach to managing high-risk AI applications. In the security domain, NATO's Principles of Responsible AI provide a relevant example of applying governance standards in defence contexts, emphasising traceability, reliability, bias mitigation, and human oversight. Building on these foundations, my research proposes an "Algorithmic R2P" framework, which seeks to extend existing international norms by addressing AI-amplified disinformation and narrative manipulation as potential triggers of real-world harm. It connects AI governance with conflict prevention by reframing such risks not only as information challenges, but as security concerns requiring coordinated response. Emerging practices such as AI red-teaming and stress-testing further complement these efforts by identifying how systems may amplify harmful narratives before they escalate. The added value of the AI Dialogue lies in its ability to connect and operationalise these approaches, aligning technical, policy, and security perspectives into a coherent and preventive global governance framework.