ANJARWALLA AND KHANNA LLP

For me, success would not just be measured by a well-worded declaration at the end of the Dialogue. It would be measured by whether what comes out of it can actually be used by people like me, working in institutions that are still figuring out what responsible AI adoption even looks like day to day. One outcome I would genuinely find meaningful is a shared baseline of standards that countries and organisations can realistically commit to, not a perfect framework, but an honest starting point that does not exclude those who are still catching up. I would also want to see developing countries treated as genuine contributors to this conversation, not just recipients of frameworks designed elsewhere. The legal and governance realities in Africa, for example, are distinct, and any dialogue that calls itself "global" has to actually reflect that. Sector-specific guidance would also matter enormously. Working in a law firm, I see firsthand how different the AI risks and questions are in legal practice compared to, say, tech or healthcare. A single sweeping framework tends to be too vague to be useful on the ground. Beyond that, I think the Dialogue needs to be honest about the fact that AI is not going to stop evolving so whatever is agreed upon has to include a real commitment to reviewing and updating the framework regularly, not just revisiting it in ten years. And finally, there has to be something for institutions with limited resources. Responsible AI adoption costs time, expertise, and money. If the outcomes do not include a practical capacity-building roadmap, then the gap between those who can afford to do this well and those who cannot will only grow wider.

Safe, secure and trustworthy AI;Transparency, accountability, and human oversight

I selected these thematic areas because they directly reflect the priorities and responsibilities of my role as a Knowledge Management Junior Officer at a law firm, as well as the broader obligations of the legal profession in relation to AI governance. Safe, secure and trustworthy AI is fundamental to legal practice. Law firms manage highly sensitive client information, and any AI system introduced into our workflows must maintain confidentiality, operate reliably, and be free from harmful errors or bias that could compromise client outcomes. Transparency, accountability, and human oversight are essential principles for responsible AI use in legal practice. Transparency requires that AI systems produce explainable outputs, enabling lawyers to understand and justify the reasoning behind AI-assisted advice or decisions. Human oversight means that as legal professionals, we must actively monitor, review, and retain the authority to override AI outputs at all times.

Although not directly stated, the thematic areas have captured data governance and confidentiality, professional liability and regulatory gaps which are great areas of concern or emerging issues.

Questions around data privacy, client confidentiality, and professional liability remain largely unanswered at a regulatory level. The Data Protection Act 2019 is a strong foundation, but it was not designed with AI specifically in mind, and that gap shows. Kenya has a young, tech-savvy population, a growing innovation ecosystem, and a government that has shown genuine interest in digital transformation. This is shown with attempts to create policies or laws around AI such as the National AI strategy 2025-2030, the AI Bill 2026. If we can build the right governance structures, ones that are practical, locally informed, and rights-based, Kenya could be a model for how African countries navigate AI responsibly.

At its best, the Dialogue can serve as a living bridge between nations that are ahead in AI development and those still laying the groundwork. Geopolitics often gets in the way of genuine cooperation, but AI governance is one area where the risks are shared enough that countries have a real incentive to work together. The Dialogue can also help move international cooperation away from declarations that gather dust toward practical mechanisms such as shared early warning systems for AI risks, mutual recognition of regulatory standards, and joint capacity-building programmes that actually reach institutions beyond the capital cities. The AI Dialogue will not solve everything but if it plants the right seeds, it could be the moment we look back on as when the world stopped talking past each other on AI and started genuinely listening.

At the international level, the UNESCO Recommendation on the Ethics of AI and the OECD AI Principles have already laid important ethical and policy groundwork. The UN Secretary-General's AI Advisory Body has produced substantive recommendations, and the Global Digital Compact has begun embedding AI governance into the broader UN framework. These are solid foundations, but they have largely operated in parallel rather than in conversation with each other. Regionally, the African Union's Continental AI Strategy is a significant initiative that reflects the specific developmental, rights-based, and infrastructure realities of the continent. For Kenya and the broader East African region, this strategy is far more grounded in local context than frameworks designed in Europe or North America. The AI Dialogue should actively engage with it rather than treat it as a footnote. In the private and professional sectors, initiatives like the Partnership on AI and various industry-led responsible AI frameworks have generated practical knowledge about implementation, the kind of on-the-ground insight that policymakers often miss. The added value the AI Dialogue can genuinely bring is coherence. Right now, institutions, governments, and practitioners are navigating a fragmented landscape of overlapping and sometimes contradictory frameworks. The Dialogue has the credibility and convening power to bring these threads together into something more coordinated without erasing the nuance that makes regional and sector-specific approaches valuable. For someone working in the legal sector in Kenya, that coherence would be transformative. It would mean clearer guidance, less duplication, and governance frameworks that are actually designed to work in practice and not just on paper.

Singapore Model AI Governance Framework - Guide for legal professionals Kenya's National AI Strategy 2025-2030