Globals ITES Private Limited

A successful Global Dialogue on AI Governance must translate vision into measurable global action. It should create a shared and interoperable framework for safe and secure AI that balances innovation with accountability, ensuring that nations do not evolve in isolated regulatory silos. From our experience at the India AI Impact Summit 2026, one of the most powerful outcomes was the convergence of defence, policy, and industry stakeholders to address real world AI risks such as autonomous decision systems, cyber warfare scenarios, and critical infrastructure protection. These discussions demonstrated that governance must be grounded in operational realities, not just policy theory. A key outcome of this dialogue should be the establishment of globally accepted safety benchmarks, including mandatory testing, auditability, and resilience against adversarial threats. AI systems deployed in sensitive sectors must be secure by design, continuously monitored, and aligned with national and global security priorities. Equally critical is the protection of human rights. AI governance must ensure privacy, fairness, and non discrimination while preventing misuse such as mass surveillance or algorithmic bias. At the India AI Impact Summit, discussions on responsible AI emphasised the importance of embedding ethics into system design rather than treating it as an afterthought. Responsible usage must also be enforced through clear accountability frameworks for developers, deployers, and governments, with human oversight remaining central in high impact systems. Ultimately, success will be defined by the ability to create trust. AI must not only be powerful, but also secure, transparent, and aligned with human dignity and global societal values.

• Safe, secure and trustworthy AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight
• Interoperability of governance approaches

Our selection reflects a deep commitment to building AI systems that are secure, accountable, and aligned with human values, based on our operational experience in cybersecurity and national critical infrastructure. At Globals, we have been recognized under the IndiaAI Mission of the Ministry of Electronics and Information Technology, Government of India, for our work in evaluating the safety and robustness of AI systems. Our focus has been on testing guard rails, identifying model biases, and assessing how AI systems, especially on premises deployments, can be abused or manipulated in real world environments. This has reinforced our belief that safe and trustworthy AI cannot be achieved without rigorous security validation and continuous monitoring. In parallel, we strongly emphasize transparency, accountability, and human oversight. In high impact systems, particularly in defence and critical sectors, human decision making must remain central, with AI serving as an assistive layer rather than an autonomous authority. Protection of human rights is equally critical. Through our work, we have observed how unchecked AI systems can lead to unintended bias, privacy violations, and disproportionate impact on vulnerable populations. This underscores the need to embed fairness, privacy, and ethical safeguards at the design stage itself. Finally, interoperability of governance approaches is essential. AI risks do not respect borders, and fragmented regulatory frameworks can create vulnerabilities. Through engagements such as the India AI Impact Summit 2026, we have actively contributed to multi stakeholder discussions that aim to harmonize global approaches to AI governance. These priorities are not theoretical for us. They are shaped by real world deployments, security testing, and a commitment to ensuring that AI evolves as a force for trust, safety, and global good.

Yes. Beyond the listed themes, one cross cutting issue that deserves greater attention is AI assurance across the full lifecycle. Governance cannot stop at principles or model release. It must also cover how models are built, fine tuned, deployed, integrated, updated, and monitored in real environments. NIST has specifically expanded secure development guidance for generative AI and dual use foundation models, recognizing that AI requires security practices throughout the development lifecycle, not only at deployment. A second emerging issue is AI supply chain integrity and model provenance. Increasingly, risk does not come only from the model itself, but from training data, external plugins, orchestration layers, fine tuning pipelines, model weights, and downstream applications. This is especially relevant for on premises and open weight deployments, where misuse, tampering, or unauthorized repurposing can be harder to detect. OECD has highlighted both the opportunities and the risks of open weight models, including privacy concerns and malicious use. A third area is content authenticity and synthetic media provenance. In an era of realistic AI generated text, audio, image, and video, governance should more explicitly address trusted provenance mechanisms for digital content. C2PA's Content Credentials framework shows the importance of cryptographically signed provenance to help verify origin and edits of digital media. These issues are highly relevant to Globals' work. Our experience in cybersecurity, adversarial testing, guard rail validation, bias assessment, and abuse testing of AI systems under the IndiaAI Mission has shown that practical AI governance must include assurance, provenance, and misuse resilience alongside ethics and human rights. Suhas Gopinath's work at the intersection of cyber resilience, critical infrastructure, and responsible AI further reinforces the need for governance that is operational, security conscious, and globally interoperable.

MThe most significant governance gap today lies in the disconnect between policy intent and operational reality. While there is growing global consensus on safe, secure, and trustworthy AI, implementation frameworks often lag behind the speed at which AI systems are being deployed, particularly in critical sectors such as defence, finance, and infrastructure. In India and across emerging economies, one key challenge is the absence of standardized mechanisms for testing, validating, and continuously monitoring AI systems in real world environments. Through our work at Globals, including recognition under the IndiaAI Mission of the Ministry of Electronics and Information Technology, we have observed that AI models, especially on premises deployments, can be vulnerable to manipulation, misuse, and unintended bias if not rigorously secured and audited. Another major gap is the lack of interoperability across governance frameworks. Fragmented regulations create compliance complexity for organizations operating globally and can inadvertently introduce security blind spots. This is particularly relevant in cyber resilience, where AI driven systems interact across borders and infrastructures. At the same time, this presents a significant opportunity. India is uniquely positioned to lead in defining practical and scalable AI governance models that integrate security, ethics, and innovation. Initiatives such as the India AI Impact Summit 2026 demonstrate how multi stakeholder collaboration can bridge the gap between policy and implementation. "AI governance must move from being a statement of intent to a system of assurance. Trust in AI will not be built by promises, but by proof." By embedding security testing, human oversight, and accountability into the lifecycle of AI systems, we can transform governance from a reactive exercise into a proactive foundation for global trust.

The AI Dialogue must evolve into a platform that transforms cooperation from alignment of principles into alignment of action. Today, AI governance is fragmented across jurisdictions, while risks such as adversarial attacks, model manipulation, and misuse are inherently global. The Dialogue can bridge this gap by enabling operational collaboration across nations. At Globals, our work across India and Europe demonstrates the importance of such cooperation. Through partnerships with CERT In, the IndiaAI Mission, Indian defence institutions, academia, AIUC, INESIA in France, IASEAI, and the HSD ecosystem in the Netherlands, we have seen firsthand that securing AI systems requires cross border knowledge sharing, joint testing approaches, and coordinated response frameworks. The Dialogue can institutionalize this by creating shared mechanisms for AI assurance, including common testing standards, red teaming frameworks, and threat intelligence exchange. It should enable trusted ecosystems where countries can collaborate without compromising sovereignty, ensuring interoperability of governance while maintaining national priorities. "AI risks are global, but governance remains local. The AI Dialogue must be the platform that closes this gap." By bringing together governments, industry, and research ecosystems, the Dialogue can accelerate the creation of secure, trustworthy, and human centric AI systems. Its true success will lie in enabling nations to not only agree on principles, but to jointly implement, validate, and enforce them in real world environments.

The AI Dialogue should build upon strong existing foundations such as the IndiaAI Mission, OECD AI Principles, G20 frameworks, UNESCO recommendations, and emerging technical standards from NIST and ISO. However, these efforts remain distributed and often lack integration at the operational level. Globals' experience across diverse ecosystems highlights this gap. Our collaborations with CERT In, IndiaAI Mission, Indian defence organizations, academia, AIUC, INESIA in France, IASEAI, and the HSD Netherlands ecosystem demonstrate the power of multi stakeholder cooperation, but also the challenges of aligning frameworks, standards, and practices across regions. The added value of the AI Dialogue lies in its ability to act as a global integrator. It can harmonize these initiatives into interoperable governance models, enable mutual recognition of standards, and create pathways for real world implementation. This includes establishing global benchmarks for AI assurance, covering security, bias, robustness, and misuse resilience. It can also strengthen public private partnerships at scale, ensuring that innovation and security evolve together. Importantly, it should empower emerging economies to participate as equal contributors, not just adopters. "AI governance does not need more principles. It needs convergence, validation, and global execution." By connecting existing initiatives and embedding them into actionable systems, the AI Dialogue can unlock trust, reduce fragmentation, and drive responsible AI adoption at a global scale.

For the AI Dialogue to be truly impactful, participation must move beyond representation to meaningful contribution. Different stakeholders bring distinct value. Governments provide policy direction and regulatory authority. Industry brings real world deployment experience and risk visibility. Academia contributes research depth and long term thinking. Civil society ensures that human rights and societal impact remain central. From Globals' experience working with CERT In, IndiaAI Mission, Indian defence institutions, academia, AIUC, INESIA in France, IASEAI, and the HSD ecosystem in the Netherlands, the most effective collaboration happens when stakeholders are not just consulted, but co creators in the governance process. The structure of the AI Dialogue should reflect this. It should include thematic working groups focused on areas such as AI security, human rights, and governance interoperability, supported by technical task forces that develop testable frameworks and standards. In addition, there should be live simulation and red teaming environments where policies are stress tested against real world scenarios. A continuous engagement model is critical. The Dialogue should not be a one time event, but an evolving platform with periodic reviews, shared intelligence mechanisms, and collaborative pilots across countries. "Participation must evolve from voice to responsibility. The future of AI governance will be shaped not by who is present, but by who is accountable." By designing the Dialogue as a system of collaboration, validation, and execution, it can ensure that diverse stakeholders contribute not just ideas, but outcomes.

A significant gap in global AI governance discussions is the underrepresentation of those who experience the real world impact of AI systems but are not part of their design or governance. This includes emerging economies, operational cybersecurity practitioners, defence and critical infrastructure operators, small and medium enterprises, and communities affected by algorithmic decisions. In many cases, governance conversations are dominated by large technology companies and policy institutions, while those dealing with implementation risks and societal consequences are underheard. From our work at Globals across national security, cyber resilience, and AI system testing, we have seen that ground level insights are essential. For example, vulnerabilities in on premises AI systems, bias in localized datasets, and misuse in constrained environments often do not surface in global policy discussions but have significant real world consequences. Inclusion must therefore be intentional and structured. The Dialogue should create dedicated platforms for practitioners, regional innovators, and affected communities to contribute insights. It should also enable participation from the Global South as equal stakeholders, not as observers. Capacity building, knowledge sharing, and access to governance frameworks must be democratized so that all nations and communities can meaningfully participate in shaping AI's future. "The most critical voices in AI governance are often the least heard. Inclusion is not a principle. It is a prerequisite for building systems that are truly fair, secure, and globally relevant." By expanding participation to those closest to impact, the Dialogue can ensure that AI governance is grounded in reality, not abstraction.

To foster meaningful and dynamic engagement, the AI Dialogue must move beyond traditional panel discussions and adopt formats that enable co creation, real world validation, and continuous collaboration. One of the most effective approaches is the use of structured innovation labs and scenario based workshops. Drawing from the experience of the World Economic Forum's Young Global Leaders community, including initiatives such as IdeaLabs and Safe AI workshops, these formats bring together diverse stakeholders to collaboratively design solutions, stress test governance frameworks, and simulate real world AI risk scenarios. Suhas Gopinath, CEO of Globals and a Young Global Leader of the World Economic Forum, has actively engaged in such platforms, which demonstrate that immersive and participatory formats drive far deeper outcomes than static discussions. The Dialogue should incorporate live red teaming exercises where AI systems and governance frameworks are tested against adversarial scenarios, including misinformation, cyber threats, and misuse in critical infrastructure. This ensures that policies are not only discussed but validated under pressure. Another impactful format is cross border policy sandboxes, where governments, industry, and academia jointly pilot governance models in controlled environments. This enables rapid learning, interoperability testing, and refinement before large scale adoption. In addition, curated small group roundtables with decision makers, practitioners, and affected communities can enable candid discussions and actionable insights that are often not possible in large forums. "Engagement must evolve from conversation to co creation. The future of AI governance will be built in rooms where ideas are tested, challenged, and proven." By integrating innovation labs, simulation environments, and continuous collaboration platforms, the AI Dialogue can become a living system that drives action, not just alignment.

Effective AI governance is emerging through a combination of regulatory frameworks, technical assurance mechanisms, and multistakeholder platforms that translate principles into practice. A strong regulatory benchmark is the European Union AI Act, which introduces a risk based approach to governing AI systems, linking obligations to safety, fundamental rights, and accountability. Complementing this, the World Economic Forum AI Governance Alliance has created a global platform that brings together governments, industry, and academia to develop actionable guidance on safe and responsible AI deployment. France has contributed through its global AI Action Summit approach, emphasizing science driven collaboration, standards, and trusted AI ecosystems. Similarly, the Netherlands has advanced practical governance through its national strategy on responsible use of generative AI, combining public sector adoption with clear safeguards on fairness, human autonomy, and transparency. India is emerging as a leader in operational AI governance. The IndiaAI Mission's focus on safe and trusted AI, including guardrails, bias evaluation, and accountability frameworks, is complemented by recent regulatory measures that mandate labeling of AI generated content, rapid removal of harmful synthetic media, and platform level accountability. This represents a decisive shift toward enforceable governance addressing real world risks such as deepfakes and misinformation. The India AI Impact Summit 2026 further demonstrated how governance can be translated into action by bringing together global stakeholders across defence, policy, and industry to address AI security, resilience, and responsible deployment. From Globals' experience working with CERT In, IndiaAI Mission, Indian defence, academia, AIUC, INESIA in France, IASEAI, and the HSD Netherlands ecosystem, the most effective approaches are those that combine policy with continuous testing, red teaming, and real world validation. The future of AI governance will be defined not by frameworks alone, but by systems that can be tested, trusted, and proven at scale.