Bharat AI Mission

A successful first Global Dialogue on AI Governance should move beyond principles to actionable global alignment, ensuring AI benefits all of humanity—especially the Global South. First, it should deliver a shared, interoperable global governance framework—not a one-size-fits-all regulation, but a set of baseline standards on safety, transparency, and accountability that countries can adapt locally. Without interoperability, fragmented policies risk widening inequalities and weakening oversight. Second, the Dialogue must establish a Global AI Capacity & Inclusion Compact, focused on democratizing access to compute, data, talent, and governance tools. For Bharat AI Mission, this is critical—AI should not remain concentrated in a few geographies but must empower "AI for a billion people." Bridging digital and governance divides is essential to equitable participation. Third, it should create a multi-stakeholder governance model in practice, not just intent—bringing together governments, startups, open-source communities, academia, and civil society. True legitimacy will come from inclusivity and co-creation. Fourth, the Dialogue should operationalize a Global AI Risk & Evaluation Framework, including shared benchmarks for frontier models, safety testing, and disclosure norms—enabling trust and cross-border collaboration. Fifth, it must embed human-centric and ethical AI principles aligned with human rights, sustainability, and societal well-being, ensuring AI augments human potential rather than exacerbating harm. Finally, success would mean committing to a time-bound roadmap with measurable outcomes—including pilot collaborations, open datasets, and governance sandboxes—ensuring the Dialogue evolves into a continuous global movement, not a one-time forum. From Bharat's lens, the ultimate success metric is clear: AI that is inclusive, open, safe, and built for global public good—not just technological advancement, but societal transformation.

• AI capacity-building
• Open-source software, open data and open AI models
• Interoperability of governance approaches
• Transparency, accountability, and human oversight

AI capacity-building is foundational-without equitable access to talent, compute, and data, the benefits of AI will remain concentrated. We believe empowering developers, researchers, startups, and public institutions is critical to democratizing AI at scale. Open-source software, open data, and open AI models are key enablers of this vision. Open ecosystems reduce entry barriers, foster innovation, enhance transparency, and allow countries to build context-aware solutions tailored to their societal needs. Interoperability of governance approaches is essential in a fragmented regulatory landscape. We advocate for globally aligned baseline standards that can be adapted locally, enabling cross-border collaboration while respecting national priorities and sovereignty. Transparency, accountability, and human oversight are critical to building trust in AI systems. As AI scales across sectors, mechanisms for explainability, auditability, and human-in-the-loop governance must be embedded to ensure responsible deployment. Together, these priorities represent a balanced approach-democratizing access, fostering innovation, enabling global cooperation, and ensuring trust. From Bharat's perspective, this is essential to realizing AI not just as a technological advancement, but as a tool for equitable societal transformation at scale.

Yes, while the identified themes are comprehensive, several cross-cutting and emerging issues require deeper attention to ensure truly inclusive and future-ready AI governance. First, equitable access to compute infrastructure is a critical gap. As AI capabilities increasingly depend on large-scale compute, there is a risk of concentration among a few countries and corporations. Democratizing access to affordable and sustainable compute must be treated as a global public good. Second, data sovereignty and data commons frameworks need clearer articulation. Countries require mechanisms to responsibly share, govern, and benefit from data while preserving privacy, cultural context, and national interests-especially for underrepresented languages and communities. Third, the rise of frontier AI models and systemic risk management calls for coordinated global approaches to evaluation, red-teaming, and incident response. This includes addressing risks such as misinformation at scale, autonomous decision-making, and concentration of power. Fourth, AI and sustainability is an emerging priority. The environmental impact of large-scale AI systems-energy consumption, water usage, and carbon footprint-must be integrated into governance frameworks. Fifth, there is a need to strengthen public digital infrastructure (PDI) for AI-including interoperable platforms, digital public goods, and open innovation ecosystems that enable scalable and inclusive deployment across sectors like healthcare, education, and agriculture. Finally, cultural and linguistic diversity in AI systems remains underrepresented. Ensuring AI reflects diverse knowledge systems, languages, and societal contexts is essential for global inclusivity. Addressing these cross-cutting issues will ensure that AI governance evolves beyond risk mitigation to enabling equitable access, shared innovation, and sustainable development, aligning with the vision of AI as a global public good. If you want, I can sharpen this further into a more diplomatic UN tone or make it more bold and thought-leadership driven.

The current AI governance landscape presents both significant gaps and transformative opportunities for India and the broader Global South. A key challenge is the asymmetry in access to compute, talent, and high-quality datasets, which risks concentrating AI innovation within a few global players. This limits the ability of emerging economies to build contextually relevant AI solutions at scale. Additionally, the absence of interoperable global governance frameworks creates regulatory fragmentation, making cross-border collaboration complex and slowing innovation. Another critical gap lies in operationalizing transparency and accountability. While principles exist, there is limited standardization around model evaluation, auditability, and explainability—especially for large, frontier models. This creates trust deficits in high-impact sectors such as healthcare, finance, and public services. From a societal lens, linguistic and cultural underrepresentation remains a major concern. For a country like India, with its vast diversity, AI systems that do not reflect local languages and contexts risk exclusion and reduced effectiveness. However, these challenges also present unique opportunities. India is well-positioned to lead through Digital Public Infrastructure (DPI) and open ecosystems, enabling scalable, inclusive AI solutions. Initiatives aligned with Bharat AI Mission can help create shared compute infrastructure, open datasets, and AI innovation platforms, lowering barriers for startups, academia, and public institutions. Furthermore, India can play a pivotal role in shaping globally interoperable yet locally adaptable governance models, balancing innovation with responsibility. By championing open-source AI, capacity-building, and human-centric governance, India can help redefine AI as a tool for equitable development and global public good.

The AI Dialogue can play a pivotal role in advancing international cooperation by evolving from a discussion platform into a mechanism for alignment, coordination, and collective action. First, it can enable the development of globally interoperable AI governance frameworks, establishing baseline standards for safety, transparency, and accountability while allowing contextual adaptation. This reduces regulatory fragmentation and fosters trust in cross-border AI systems. Second, the Dialogue can serve as a catalyst for inclusive global participation, ensuring that voices from the Global South are actively represented in shaping AI norms. This is essential to avoid concentration of power and to ensure AI governance reflects diverse socio-economic and cultural realities. Third, it can drive collaborative capacity-building initiatives—including shared access to compute infrastructure, datasets, and talent development programs. Such cooperation can significantly reduce capability gaps between developed and emerging economies. Fourth, the Dialogue can promote open innovation ecosystems, encouraging the adoption of open-source AI models, open data frameworks, and digital public goods. This fosters transparency, accelerates innovation, and enables scalable solutions for global challenges. Fifth, it can facilitate the creation of joint mechanisms for risk assessment and incident response, including shared evaluation benchmarks for frontier AI systems. Coordinated approaches to safety will be critical as AI capabilities rapidly evolve. Finally, the Dialogue should institutionalize multi-stakeholder collaboration, bringing together governments, industry, academia, and civil society to co-create solutions and ensure accountability. From Bharat AI Mission's perspective, the Dialogue's true value lies in its ability to transform cooperation into co-creation—building an AI ecosystem that is inclusive, open, secure, and aligned with the vision of AI as a global public good.

The AI Dialogue should build on and connect with existing global, regional, and multi-stakeholder initiatives to avoid duplication and accelerate convergence. Key global efforts include the United Nations Global Digital Compact, OECD AI Principles, GPAI (Global Partnership on AI), and UNESCO's Recommendation on the Ethics of AI, all of which provide foundational norms on responsible and human-centric AI. Additionally, emerging safety-focused platforms such as the Bletchley Park AI Safety Summit outcomes and frontier model evaluation initiatives offer important building blocks for risk governance. From an infrastructure and inclusion lens, initiatives around Digital Public Infrastructure (DPI)—such as India Stack—along with open-source ecosystems and data commons movements, provide scalable models for inclusive AI development. Regional collaborations and South-South partnerships are also critical to ensure context-aware innovation. The added value of the AI Dialogue lies in its ability to act as a convergence layer—bridging fragmented efforts into a more coordinated and interoperable global ecosystem. It can harmonize standards, align definitions, and facilitate mutual recognition of governance frameworks across jurisdictions. Further, the Dialogue can operationalize collaboration by enabling joint pilot projects, shared compute and data resources, and cross-border sandboxes for testing AI systems in real-world contexts. It can also serve as a platform to amplify the role of open-source communities, startups, and academia, which are often underrepresented in global policy forums. Importantly, the Dialogue can strengthen Global South participation, ensuring that governance frameworks are not only globally accepted but also equitable and inclusive. From Bharat's perspective, the Dialogue's value lies in transforming existing principles into coordinated action, shared infrastructure, and measurable outcomes, advancing AI as a true global public good.

Inclusive participation is essential for the AI Dialogue to be legitimate, impactful, and future-ready. The Dialogue should adopt a multi-stakeholder, multi-layered, and action-oriented structure that enables meaningful contributions from all actors. Governments should provide policy direction, enable regulatory alignment, and support public infrastructure for AI. Industry and startups can contribute through innovation, real-world deployment insights, and responsible AI practices. Academia and research institutions should anchor technical standards, evaluation frameworks, and capacity-building initiatives. Civil society and communities must play a central role in ensuring inclusivity, ethics, and representation of diverse societal needs, particularly from the Global South. To operationalize this, the Dialogue should be structured around thematic working groups (e.g., safety, open ecosystems, capacity-building, governance interoperability), each co-led by representatives from different stakeholder groups and geographies. These groups should focus on time-bound deliverables, such as policy toolkits, open datasets, or pilot programs. Additionally, the Dialogue should adopt a hybrid and decentralized format, combining global convenings with regional and local consultations. This ensures broader participation, especially from underrepresented regions and grassroots innovators. A digital collaboration platform should be established to enable continuous engagement, knowledge sharing, and co-creation beyond annual events. This can include open repositories, discussion forums, and progress dashboards. Importantly, participation must go beyond representation to shared ownership—through mechanisms such as co-creation labs, innovation challenges, and cross-border sandboxes. From Bharat's perspective, the Dialogue should embody the principle of "AI by the world, for the world"—ensuring that diverse voices not only contribute but actively shape outcomes, driving inclusive and responsible AI for global public good.

Global discussions on AI governance continue to underrepresent several critical voices, particularly from the Global South, grassroots innovators, and non-urban communities. Countries with emerging digital ecosystems often lack proportional influence in shaping norms, despite being among the largest future users of AI. Linguistically and culturally diverse communities are also underrepresented. Much of AI development and governance is centered around a limited set of languages and contexts, which risks excluding billions of people whose realities are not reflected in current systems. Startups, open-source communities, and independent researchers—key drivers of innovation—often have limited access to global policy platforms compared to large technology corporations. Similarly, civil society organizations, especially those working at the intersection of AI and social impact, are not consistently included in decision-making processes. Additionally, youth voices and future generations, who will be the most affected by AI, are rarely meaningfully engaged in governance dialogues. To address this, the AI Dialogue should adopt deliberate inclusion mechanisms. This includes ensuring geographic and linguistic diversity quotas, funding participation from low- and middle-income countries, and enabling regional and grassroots consultations that feed into global discussions. Creating open digital participation platforms can allow broader contributions beyond physical convenings. The Dialogue should also institutionalize roles for open-source communities, startups, and civil society within working groups and decision-making structures. Further, promoting multilingual AI governance resources and datasets will ensure more inclusive engagement and context-aware policymaking. From Bharat AI Mission's perspective, inclusion must move from representation to co-creation—ensuring that underrepresented voices are not just heard, but actively shape the frameworks that will govern AI globally. This is essential to building AI that is equitable, culturally relevant, and truly global.

To foster meaningful and dynamic engagement, the AI Dialogue should move beyond traditional panel discussions and adopt participatory, action-driven, and technology-enabled formats. First, co-creation labs and policy sprints can bring together governments, industry, academia, and civil society to collaboratively design solutions—such as governance toolkits, safety benchmarks, or open datasets—within defined timeframes. This ensures outcomes are tangible and implementation-ready. Second, AI sandboxes and live use-case demonstrations can enable stakeholders to test governance approaches in real-world scenarios across sectors like healthcare, agriculture, and education. This bridges the gap between policy and practice. Third, global innovation challenges and hackathons focused on themes like responsible AI, multilingual models, or public good applications can crowdsource solutions from startups, developers, and students worldwide, especially from underrepresented regions. Fourth, the Dialogue should incorporate regional dialogues and decentralized forums, feeding into the global platform. This ensures context-specific insights and broader participation, particularly from the Global South. Fifth, a persistent digital collaboration platform should complement physical convenings—featuring open repositories, discussion forums, working group dashboards, and multilingual participation tools—enabling continuous engagement and transparency. Additionally, formats like reverse panels (where policymakers listen to grassroots innovators), citizen assemblies, and youth councils can bring fresh perspectives into governance discussions. Finally, embedding time-bound deliverables and public progress tracking will ensure accountability and sustained momentum. From Bharat' AI Missions's perspective, the Dialogue should embody a shift from conversation to co-creation at scale, leveraging innovative formats to build inclusive, actionable, and globally relevant AI governance outcomes.

Several emerging policies and practices offer valuable models for effective and inclusive AI governance. India's approach to Digital Public Infrastructure (DPI)-such as Aadhaar, UPI, and ONDC-demonstrates how open, interoperable, and population-scale platforms can enable innovation while maintaining governance oversight. Extending this model to AI through shared compute, open datasets, and public platforms can democratize access and accelerate adoption. The OECD AI Principles and UNESCO Recommendation on the Ethics of AI provide globally recognized frameworks for human-centric, transparent, and accountable AI, helping align policy across jurisdictions. The EU AI Act represents a risk-based regulatory approach, categorizing AI systems based on potential harm and introducing proportionate compliance requirements. This offers a structured model for balancing innovation with safety. Open-source AI ecosystems-including collaborative platforms for models, datasets, and evaluation-serve as powerful enablers of transparency, innovation, and inclusivity. They lower barriers to entry and allow countries to build context-aware solutions. Regulatory sandboxes are another effective practice, enabling controlled experimentation with AI systems in sectors like finance and healthcare. These allow policymakers and innovators to co-develop regulations based on real-world insights. Additionally, initiatives focused on AI safety and evaluation, including shared benchmarks and red-teaming exercises, are critical for managing risks associated with advanced AI systems. Finally, multilingual AI development efforts and data commons initiatives help address linguistic and cultural gaps, ensuring broader representation and equitable access. From Bharat AI Mission's perspective, the most effective approach combines open ecosystems, interoperable governance, and scalable public infrastructure, enabling AI to be developed and deployed as a global public good-secure, inclusive, and innovation-friendly.