AI synergy: Letting local realities guide global AI innovation toward impact

Artificial Intelligence (AI) may be global in its reach—but its impact is deeply local.

While groundbreaking innovations in AI – broadly defined - still originate predominantly from a handful of powerful labs in the world’s richest countries, the most profound social and economic effects of AI won’t be felt in Silicon Valley — they will materialize in cities like Accra, Amman, Karachi, Lima, Lome, Rabat, Vientiane, or Suva.

AI sub-fields like prediction and decision making, machine vision and sensing, natural language processing and speech recognition, robotics and automation, and generative AI offer immense potential. They’re accelerating discoveries in climate science, and improving outcomes in agriculture, health, and education. But their impact in low- and middle-income countries hinges on something more: local relevance, local data, and local agency.

To unlock AI’s transformative power everywhere, we must connect the best of global technology with the lived realities of communities. That means improving access to connectivity and energy, compute and models, context - in the form of high quality, well governed data, capabilities to develop and use AI solutions and developing AI use cases that reflect cultural context and community priorities. To move fast and at scale, we must pair AI solutions that can enhance development outcomes within existing constraints with efforts that build the ecosystem and investment appetite needed for developing economies to become AI innovators themselves.

At the World Bank Group’s Global Digital Summit 2025, our panel of speakers explored this intersection—how to avoid false choices between “global” and “local,” and instead create synergies that drive inclusive innovation. Three themes emerged that can help chart this middle path.

First, focus on local data to unlock local benefits: While powerful ML tools, including LLMs, are driving remarkable progress in fields like education, healthcare, and agriculture, they often fall short of their potential when applied to specific local challenges in emerging economies without adaptation. Communities could benefit significantly more from customized AI solutions developed or fine-tuned for their contexts, using data that reflects their needs, values, and preferences.

Across Africa, the resurgence of small-scale, highly specialized AI models – so called ‘small AI’ - reflects a deeper, more effective understanding of regional needs and resources. Enabling cross-border data flows and connecting global research outfits that build the models to local research centers that can bring the local context can be effective. The Togo Data Lab, a partnership between the Togo’s Ministry of Digital Economy and Transformation and University of California, Berkeley, aiming to establish a sustainable data science function within the Togolese government is a useful example.

Second, strengthen digital infrastructure and education to localize innovation: Reliable foundations for AI, including connectivity and sustainable energy, computing provided either on the edge, in data centers, or via the cloud, context in the form of high quality, well governed data, and capabilities to develop use AI are indispensable. To help benefits of AI reach vulnerable groups, targeted educational initiatives should empower diverse populations, including historically marginalized groups, to actively participate in AI innovation.

Programs that provide training and resources specifically to underrepresented groups have already shown their effectiveness in driving innovative, locally relevant solutions even without extensive computational power or vast datasets. An example is the African Women in AI and Tech program run in partnership between UNESCO, the International Artificial Intelligence center of Morocco, and the OCP Foundation. Moreover, nurturing smaller, targeted AI innovations that address local problems, rather than exclusively pursuing massive, commercially driven models is essential. Such tailored AI applications, supported by robust data science, have a greater potential for sustainable, community-focused solutions that meet real-world needs. Finally, countries need to explore how they can leverage their assets – be it mineral resources or data – more effectively to finance investments that allow them to move from AI consumers to AI producers.

Lastly, shape AI governance that balances international cooperation with national sovereignty: Effective AI governance does not mean uniformity; rather, it involves creating standards and frameworks that support collaboration without compromising a country's autonomy or control over resources. Developing clear, governance models that enable countries to collaborate and share as peers and that prioritize interoperability can strengthen global collaboration while preserving sovereignty. Starting with pilot programs among a small group of countries to test and refine these models and enforceable rules at the regional level could help establish valuable precedents for wider application. 

Key Actions to Unlock Synergies Between Local and Global AI

The conversation at the Global Digital Summit 2025 made one thing clear: the way forward is not global or local, but a middle path that intertwines aspects of both approaches to bring AI-enabled solutions and innovation to the community level in developing economies. On the one hand, global models and infrastructure will continue to power breakthroughs in AI capabilities — but without local context, and adaptation they can’t unleash their full potential. On the other, local, ‘small AI’ models based on specific data sets can more effectively and efficiently solve many of today’s issues, without significant investment. Local actors hold the insights, data, and community trust needed to ensure that AI – in all its forms - is not only transformative, but trustworthy and reliable.

The panelists pointed out Six Imperatives for Action:

• Invest in local and diaspora talent to fuel inclusive innovation grounded in national priorities and regional contexts.
• Build from within by strengthening education, infrastructure, and capacity before over-relying on imported solutions.
• Engage communities in decision-making processes to ensure that AI development is democratic, inclusive, and aligned with the lived experiences of those it aims to serve.
• Prioritize real-world use cases that directly address local challenges and create meaningful value for citizens, especially in underserved areas.
• Adopt a proactive governance stance, among policymakers and institutions, recognizing AI’s potential and acting with purpose to shape its local applications.
• Think beyond the short-term to build sovereign, sustainable, and context aware AI ecosystems.

Entrepreneurs, private sector, academia, and governments in developing economies are already thinking global and acting local to bring relevant and effective AI solutions to the contexts where they can make the biggest difference. Development institutions and other partners should help them scale up these efforts to unlock AI for development.