Artificial Intelligence governance was a central focus of discussions at a high-level roundtable co-convened by UNDP and GSMA during Mobile World Congress 2026 in Barcelona in March. Bringing together ministers, mobile industry leaders, academics, civil society representatives, and international organizations, the discussion focused on governance approaches that can support inclusive, practical AI deployment.  

One message stood out clearly: the mobile industry is the AI gateway for billions. 

It is easy to talk about AI in the language of hyperscalers and frontier models. But for many people around the world, AI will likely arrive through a smartphone. Across the Arab States and much of the Global South, people are not laptop-first or desktop-first - they are mobile-first. The handset is the primary delivery platform for services, information, financial access, and increasingly, AI-powered tools. 

Connectivity is a critical part of that foundation. But beyond  infrastructure, the mobile sector is actively driving AI exploration: catalyzing startups, enabling edge computing, developing new business models, and serving as a powerful  channel for talent development at scale. If AI is to reach underserved communities, mobile is not just a pathway – it is the pathway. 

No single actor can shape the AI era 

One of the roundtable's most important  insights was also one of its simplest: no single country, sector, or institution can shape the AI era alone. The complexity of what is being built, and the speed at which it is moving, demand new forms of collaboration. 

Participants pointed to the quadruple helix model as a promising framework: bringing together government, private sector, academia, and civil society not as separate stakeholders providing input, but as co-designers of AI ecosystems.  

Governments, in particular, were encouraged to think beyond their borders, engaging diaspora communities living and working in global AI hubs as a source of talent, insight, and international connection that national strategies too often overlook.  

Infrastructure is not optional  

A critical reminder emerged throughout the discussions:  national AI strategies are not just about AI.  

Reliable and affordable power supply, broadband infrastructure, interoperable digital systems, these foundational layers determine whether AI can function at all. Governance frameworks that ignore physical and digital infrastructure are, quite simply, building on sand.  

This is especially relevant in emerging contexts, where gaps in infrastructure risk becoming bottlenecks to AI adoption. Strengthening these foundations is not separate from AI governance - it is central to it. 

Agility matters - but it is not enough  

Technology is evolving rapidly, and citizen expectations are evolving faster than many regulatory systems can adapt. Bridging that gap is one of the central governance challenges of the AI era.  

Regulatory sandboxes were discussed as a valuable tool, creating space for experimentation, building trust between regulators and innovators, and helping to shape more dynamic institutional cultures. However, the discussion also underscored the limits. Sandboxes are not a substitute for strong underlying policy. Short-term pilots cannot compensate for weak, incoherent, or outdated regulation. 

The goal is not just agility for its own sake, it is building governance systems that are both adaptive and durable.  

As one participant noted through an Arab proverb: the wind does not break a tree that bends.  

AI is a layer, not a reset  

In countries still strengthening their digital foundations, AI is sometimes presented as an opportunity to leapfrog  digital transformation. However, the roundtable emphasized a more grounded reality: AI deployment  depends on what already exists in terms of  digital infrastructure and governance systems.  

AI is not a replacement for existing digital infrastructure. It is a layer built on top of it.  

Digital Public Infrastructure (DPI) – including  data exchange mechanisms, canonical registries, interoperable identity systems - remains foundational. The hard-won lessons of previous digital eras, including investment in digital skills, privacy regulations, and human-centered design practices, are not irrelevant in the AI age. They are amplifiers.  

Countries should build on what exists, not start from scratch. Those that have invested in DPI and digital governance capacity are better positioned to deploy AI responsibly and at scale. Those that have not must ensure that AI efforts  do not reinforce existing inequalities.  

Inclusion is not a side agenda, it is the agenda  

A key takeaway was clear: the risks of AI are not only about what the technology can do, but what and who it serves.   

Who has access? 

Who shapes it? 

Whose realities are reflected in the design?  

Countries risk entrenching existing digital divides or creating new ones if inclusion is treated as a downstream consideration rather than a design principle. More voices at the table do not just make AI more equitable. They make it more innovative. Perspectives from underrepresented communities, languages, and contexts surface problems and possibilities that homogeneous development and deployment environments simply miss.  

The AI era may not be defined solely by who can build the largest models. It may equally be shaped by who can build the most contextually relevant ones: small-language models, sector-specific tools, and locally grounded applications. For the Arab States, this competitive and developmental opportunity.  

The playbook is not written and that is the point.  

The roundtable closed with a clear shared understanding: no country or region has this figured out. Every country, every institution, every policymaker is navigating this in real time. The frameworks are incomplete. The evidence base is still forming. The standards are still being shaped.  

But this uncertainty is not a reason to wait – it is a reason to act intentionally and collaboratively now. The standards, norms, and governance architectures being built today will shape who benefits from AI for decades. Countries that invest early in governance capacity, digital infrastructure, and multi-stakeholder collaboration are likely to be better positioned to shape and benefit from AI adoption. The UNDP Regional Bureau for the Arab States (RBAS) is working with countries across the region to leverage this opportunity. The forthcoming UNDP RBAS AI Deployment Framework  will provide a pathway for countries in defining, shaping, and accelerating their AI journeys.  

UNDP welcomes  partnerships - with public sector, private sector, and civil society AI decision-makers, policymakers, and innovators - to advance AI efforts across the Arab States. 

 For partnership inquiries, contact 

Dany Wazen, Head of the Regional Digital and AI Team, UNDP Regional Bureau for Arab States: dany.wazen@undp.org