DataConfo

The most meaningful outcome would be the establishment of a shared framework for regulatory interoperability, one that allows national and regional AI governance regimes to recognize and align with each other without requiring uniform adoption of a single model. From a practitioner's standpoint, the current fragmentation creates real compliance barriers, especially for organizations operating across jurisdictions with different legal traditions. A successful Dialogue would produce concrete mapping mechanisms between existing frameworks (EU AI Act, African Union frameworks, national data protection laws such as Morocco's Law 09-08), agreed minimum standards for AI accountability and transparency, and inclusive technical assistance pathways for countries still building their regulatory capacity. Success is not a declaration, it is operational convergence.

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

These four themes are interdependent in practice. Safe and trustworthy AI is the baseline ,without it, governance becomes reactive rather than preventive. Interoperability of governance approaches ensures that regulatory efforts across jurisdictions reinforce rather than contradict each other, reducing compliance fragmentation for organizations and governments alike. Human rights protection must remain the non-negotiable anchor of any governance architecture, particularly as AI systems increasingly mediate access to services, justice, and opportunity. Finally, transparency and human oversight are the operational mechanisms that make the other three principles enforceable. From a field perspective, frameworks that lack clear oversight mechanisms remain aspirational,they do not translate into accountability.

One critical gap is the question of data sovereignty and sovereign AI. Current thematic areas focus on governance of AI systems, but do not sufficiently address who controls the data infrastructures on which AI is built and deployed. For many countries in the Global South, AI dependency is not primarily a question of regulation, it is a structural question of where data is stored, processed, and monetized. Without addressing data sovereignty, AI governance frameworks risk legitimizing extractive data flows under a compliance veneer. A meaningful Global Dialogue must incorporate binding principles around data localization rights, equitable access to AI infrastructure, and the right of states to develop sovereign AI capabilities without being constrained by proprietary ecosystems dominated by a small number of jurisdictions and corporations.

In Morocco, the primary data protection framework, Law 09-08, was enacted in 2009, predating modern AI systems entirely. It does not address automated decision-making, algorithmic accountability, or AI-specific data flows. Organizations are deploying AI tools while regulators operate with instruments that were never designed for this reality. This gap is not a Moroccan exception, it is the norm across the Global South. The most significant challenge is governance by default: in the absence of local AI frameworks, multinational technology providers effectively set the standard through their own contractual terms, with limited accountability to national authorities or affected populations. Data generated locally is processed, monetized, and governed elsewhere. Internationally, the fragmentation of AI governance regimes, EU AI Act, US executive orders, emerging African Union frameworks ,creates compliance complexity that smaller economies and organizations cannot absorb alone. Without interoperability mechanisms, countries like Morocco face a forced choice between adopting foreign frameworks wholesale or remaining ungoverned. The opportunity is real, however. Morocco's position at the intersection of African, Arab, and European regulatory spheres makes it a natural candidate for a regional AI governance hub , provided the Global Dialogue produces actionable alignment tools, not only declarations. Countries that build AI-ready regulatory infrastructure now will attract responsible investment and shape continental norms. Those that wait will inherit frameworks designed without them.

The AI Dialogue can serve a function that no existing multilateral forum has fully assumed: translating political consensus on AI principles into operational cooperation mechanisms. Declarations exist. What is missing is the architecture to implement them across jurisdictions with unequal resources and divergent legal traditions. Concretely, the Dialogue can play three roles. First, as a mapping platform , systematically documenting existing national and regional AI governance frameworks to identify convergence points and interoperability pathways, rather than starting from a blank slate at every negotiation. Second, as a capacity bridge, connecting countries that have regulatory frameworks but lack implementation expertise with those developing capacity but lacking legal models. Third, as a legitimacy mechanism, ensuring that governance standards emerging from this process carry genuine multilateral backing, rather than reflecting the priorities of a small number of technologically dominant states. From a practitioner's perspective, the most valuable outcome would be a standing technical working group with real participation from the Global South, tasked with producing practical interoperability tools: mutual recognition templates, cross-border incident reporting protocols, and shared definitions for high-risk AI applications. The Dialogue should not produce another framework. It should make existing frameworks work together.

Several initiatives provide a foundation the Dialogue should actively connect with rather than duplicate. The African Union's Continental AI Strategy offers a regional governance vision that remains under-resourced and insufficiently integrated into global conversations. The UNESCO Recommendation on the Ethics of AI (2021) provides an agreed ethical baseline across 193 member states that should serve as the normative anchor for the Dialogue. The OECD AI Principles and the Global Partnership on AI (GPAI) have produced useful technical work but remain concentrated among high-income economies. The added value the AI Dialogue can bring is inclusion by design, ensuring that frameworks developed through these existing initiatives are stress-tested against the realities of countries with limited regulatory infrastructure, limited compute sovereignty, and legal systems not originally designed for digital governance. Morocco's experience operationalizing Law 09-08 in an AI environment, for instance, surfaces practical gaps that purely technical or high-income-country-centric processes do not capture. The Dialogue should institutionalize this kind of ground-level feedback as a standing input, not an afterthought.

The Dialogue must move beyond panel formats. Governments bring legitimacy but should come with concrete regulatory experience. Civil society brings accountability and ground-level impact data. Private sector brings technical expertise but must operate under conflict-of-interest safeguards. Academic and practitioner communities, including from the Global South, bring implementation knowledge that is systematically absent from high-level forums. Structure recommendation: working groups by thematic area, with mandatory Global South co-chairs and binding output commitments, not advisory opinions.

Francophone Africa, Arab states, and Small Island Developing States are structurally absent from AI governance conversations despite being primary recipients of AI systems they did not design and cannot yet regulate. Inclusion requires more than translation, it requires shifting where expertise is recognized. Practical steps: regional pre-consultations feeding into the global process, funded participation for practitioners from underrepresented regions, and recognition of non-English regulatory frameworks as legitimate reference points.

Three formats would make a concrete difference. First, regulatory simulation exercises, participants stress-test draft governance tools against real national contexts, exposing gaps before adoption. Second, a live case registry ,documented AI incidents from underrepresented regions, used as evidence base during deliberations rather than anecdote. Third, asynchronous structured consultation prior to in-person sessions, so participation is not gated by travel budgets. Dialogue outputs should be versioned and publicly trackable, so commitments do not disappear between sessions.

Three approaches are worth highlighting. Morocco's Law 09-08 proves that data protection enforcement is achievable in middle-income contexts, its declaration-based mechanism and supervisory authority model are directly adaptable for AI oversight. The EU AI Act's risk-based tiered structure offers a replicable template that proportionates compliance obligations to actual risk, making it accessible beyond high-income jurisdictions. Canada's Directive on Automated Decision-Making demonstrates that algorithmic impact assessments can be implemented incrementally, without waiting for a fully mature regulatory framework. The common thread: proportionality and adaptability. The Dialogue should systematically curate such models with implementation guidance calibrated to different regulatory maturity levels - not just document best practices, but make them transferable.