By Aubra Anthony, Oluwaseun Adepoju, Aalia Meera Garrett, Megan Roberts, Lakshmee Sharma, Vitalii Zakhozhyi

Artificial intelligence (AI) is already transforming many aspects of life and has the potential to accelerate development. However, today AI’s benefits are not evenly distributed, and its rapid expansion is also introducing new challenges and exacerbating existing ones, including deepening digital divides, reinforcing biases, automating digital threats, and further concentrating technological and geopolitical power. Realizing the potential benefits of AI requires more globally equitable distribution and an inclusive approach to developing the frameworks that guide its use. Promising initiatives to do so are underway across Global Majority countries. It is critical that the insights and good practices from these experiences inform next steps in the global AI transformation.

This was the inspiration behind a recent panel discussion organized by the Carnegie Endowment for International Peace, the Co-Creation Hub, and the United Nations Development Programme (UNDP) during the 20th annual International Governance Forum held in Lillestrøm, Norway in June. The discussion included leading experts from government, civil society, the private sector, and international organizations, and focused on surfacing insights from AI applications already in use to accelerate development.

The discussion identified three crucial insights.

First, the widening AI equity gap is already exacerbating inequality and undermining development, especially in the Global South. 

The widening AI equity gap is a growing concern that threatens to deepen existing global inequalities, particularly in the Global South. For example, at present, only 2 percent of global data centers are located in Africa and just 5 percent of African AI innovators have access to necessary computational resources. 

“While we’re tremendously optimistic about the role that AI can play in supporting sustainable development, we’re also conscious of the significant AI equity gap between the Global South and the Global North - and the question of how many people will potentially be left behind in this global AI revolution.”

     - Yu Ping Chan, Head of Digital Partnerships and Engagements, UNDP Digital, AI, and Innovation Hub

This striking imbalance highlights the structural barriers that exclude many countries and communities from fully benefiting from AI-driven advancements. Without urgent and intentional interventions to address these disparities, AI risks amplifying, rather than alleviating, inequality, undermining sustainable development efforts in areas where they are needed most. 

“We need good evidence of where the impact [of AI] is - what’s working, and what are the main issues related to AI… Of course, there are cases in which the technology is not being used for the best purposes, but we also see that there are some benefits. That’s something we want Global South countries to understand - what works and what doesn’t - so they can better deploy these technologies… We need to analyze more of the implementation side - what is working, and what is not. That will help us be more efficient with the resources we have.”

     - Armando Guio Español, Executive Director, Global Network of Internet & Society Centers

The alarming reality of this notion was underscored through both panel discussions and audience engagement. When asked about optimism regarding AI’s ability to accelerate inclusive sustainable development within the next five years, the audience responses showed a notable, and nearly even split, with some participants expecting AI to accelerate development for all and others concerned about the inevitable challenges ahead. This backdrop establishes the need for deliberate policies, investments, and collaborative approaches that close the AI equity gap, ensuring that the benefits of AI technologies are widely shared, rather than concentrated among a few economic powerhouses or regions.

“Power is becoming incredibly concentrated - with just a few multinational players dominating the discourse, dominating the priority setting, and dominating the types of business models that end up getting pushed out. And often these business models aren't serving the populations that are most in question… There’s a perceived urgency to catch-up, especially in Africa, where projections show that GDP growth attributed to AI may be 10 times lower than AI-fueled growth elsewhere.” 

     - Aubra Anthony, Nonresident Scholar, Carnegie Endowment for International Peace

Second, AI systems must be human-centered, inclusive, and embedded in local contexts. 

AI systems must be designed with a human-centered approach that prioritizes inclusivity and local relevance to truly foster sustainable development. As emphasized during the forum, technology development cannot simply be imported or imposed “from above” or from afar. Instead, AI ecosystems need to be co-created with the communities they serve, drawing on local languages, cultures, and socio-economic realities. This approach prioritizes reinforcement rather than extraction, empowering local populations not only as data providers, but as active participants in defining challenges and crafting solutions that reflect their true needs. 

“For AI to be local and to be impactful in the grassroots, linguistic equity is very important… AI must be built by local people, and people must feel they co-created the solution they are using… We’ve seen situations where people invested millions of dollars in tech solutions that communities rejected - because they weren’t involved. Local doesn’t mean just technical - it means people contributing to the design, validation, and feedback of AI tools meant for them.”

     - Oluwaseun Adepoju, Managing Partner, Co-creation HUB (CcHUB)

As noted by panelists, at present many communities in the Global South often contribute the underlying data essential for AI, but have limited channels available for inputting into the governance, ethical considerations, or design of AI applications. Shifting this paradigm toward inclusive, context-aware development must be a greater priority for global AI governance.

Third, AI must be deployed sustainably and in ways that center, rather than sideline, communities and their concerns. 

Deploying AI sustainability and in ways that prioritize community needs and concerns is essential for equitable development. Environmental impacts, such as the energy or water consumption of large-scale AI models, pose ethical questions about balancing innovation with climate responsibility.

“We have to see where we balance the SDG objectives for renewable energies and climate with more efficient computing systems… The benefits of AI applications should not outweigh the costs that come from high energy usage.” 

     - Abhishek Singh, Additional Secretary, Ministry of Electronics and Information Technology, Government of India

Speakers and participants alike stressed that sustainability should include social and governance dimensions, not just technical efficiency. 

“Two big strands are emerging in this space: one is sustainable AI itself - how do we make AI more local, more clean, more green, more safe, more private, more fast, more cheap?... The second is what AI can do for larger sustainability. So it's not a technology conversation, it's truly a developmental conversation.”

     - Anshul Sonak, Principal Engineer & Global Director, Intel Digital Readiness Programs

Moreover, communities must be centered in AI governance frameworks to ensure transparency, fairness, and accountability. 

For instance, one Slido question asked the audience to name the #1 priority for enabling an inclusive AI ecosystem, with responses frequently pointing to community empowerment, participatory governance, and local capacity development.