Missed GEM25? Watch the main panel discussions on-demand and discover how AI is transforming global development.
By Mary Mei
The Harvard Center for International Development (CID) held its 16th annual Global Empowerment Meeting (GEM), GEM25: Catalyzing AI for Inclusive Change, on April 29, 2025, at Harvard Business School. Cohosted by the Digital Data Design Institute at Harvard (D^3), GEM25 focused on both the promising power of artificial intelligence (AI) in advancing development and its potential to widen existing inequities in emerging economies.
The conference emphasized the intersection of innovation and responsibility, exploring pathways for how AI can create inclusive growth. Central to GEM25’s discussions were the practical applications of AI across development sectors and the roles of governments, the private sector, and individuals in shaping and managing those applications.
GEM25 convened a diverse crowd of leaders and experts from across the globe for a full day of dialogue and discussion, including two panels, a guided AI exercise, and five interactive GEM Incubation Rooms. These Incubation Rooms allowed participants to collaborate and brainstorm solutions to issues in areas ranging from AI in education to industrial policy and AI.
Unlocking AI's Potential for Emerging Economies
On April 28, 2025, GEM opened with a private dinner featuring distinguished guests from the Harvard Ministerial Leadership Program including ministers of finance, economy, and planning from across Africa and Nepal and remarks from Luis Franceschi, Assistant Secretary General of the Commonwealth of Nations and Ronnie Chatterji, Chief Economist at OpenAI.
Instead of attending in person, Franceschi appeared as a digital avatar to lead a simulated AI-powered cabinet meeting with Kenyan avatar ministers, demonstrating how AI can support public policy decision-making in tackling issues like tuberculosis.
Chatterji highlighted successful use cases of AI and the global optimism behind its potential, particularly in Asia and Africa. He underscored the importance of identifying suitable problems for AI application and involving the right actors to guide implementation—themes that framed the next day’s discussions.
Bridging the AI Knowledge Gap: Opening Reflections from GEM25 on Leadership, Policy, and Responsibility
In his opening reflection at GEM25, Karim Lakhani, Faculty Director of D^3, raised a key concern: are leaders in government and the public policy arena ready for the dramatic impact of generative AI? Lakhani emphasized closing what he identified as the “knowing-doing” gap—bridging knowledge and practical application of AI to answer public policy questions.
Lakhani also urged participants to take a role in steering AI’s trajectory. “We should not sit there as bystanders, seeing this all unfold. We should be activists in our work to make sure that these technologies deliver what we think they should—for our countries and for our people.”
“We should not sit there as bystanders, seeing this all unfold. We should be activists in our work to make sure that these technologies deliver what we think they should—for our countries and for our people.”
Karim Lakhani, Founding Chair of Digital Data Design Institute at Harvard
Following Lakhani, Harvard Kennedy School Dean Jeremy Weinstein recognized the need to balance enthusiasm for AI with the understanding of its consequences, noting, “Even in the face of so much promising change, we must maintain a sense of sobriety and pragmatism, because so-called 'AI for-good interventions' can reinforce disparities.”
Asim I. Khwaja, CID Faculty Director, closed the opening reflections by highlighting the power of collaboration and reimagination. “This is a moment in development, in the world, in academia, to reimagine what we can do. What I think is spectacular about AI is it can allow us to reimagine the world in a completely different way,” Khwaja said.
Contextualizing AI in Emerging Economies: Ground-Level Perspectives from GEM25’s Opening Panel
Moderated by Han Sheng Chia, Policy Fellow at the Center for Global Development, GEM25’s first panel, What do Emerging Economies Want from AI?, explored the development and use of AI in emerging economies, as well as the challenges and opportunities associated with proper deployment.
The panelists offered their own perspectives on the optimism behind AI in Africa, with some questioning the context surrounding that optimism and others highlighting the resource limitations of countries in implementing general-use AI.
Umar Saif, Founder & CEO of aiSight.ai, cautioned against oversimplified enthusiasm and emphasized the need to consider tradeoffs. “If you wanted to use Chat GPT as a tutor in a country like Pakistan, for the cost that you would incur in a month, you could hire 17 full-time school teachers,” Saif explained. “It’s not the question of cost but also opportunity cost. What else can you do with those resources?”
The panel also reinforced importance of aligning AI solutions with local and societal contexts. Nanjira Sambuli, a scholar at The Carnegie Endowment, identified the pitfalls of applying generic AI models. “The kind of advice folks in development tend to think entrepreneurs or people in low resourced environments want—it tends to be condescending to be honest. [For example], “think about this cookie cutter model of things,” and actually I’m trying to make sense of it in my context,” Sambuli remarked. “Is that context studied in the code that is supposed to be feeding the AI model? Most times, not. We have a saying in Kenya: on the ground, things are different.”
Shikoh Gitau, CEO of Qhala, echoed the need for culturally aware AI development. “Our identity is formed within our community, and the moment you lose that as you’re building your platform, then there is a problem,” Gitau said. She also noted that “raw encoded intelligence,” should be integrated with human insights, which bring community intelligence and ancestral knowledge.
Teresa Clarke, Chairman & CEO of Africa.com, underscored the parallels between AI and traditional wisdom. “We think of AI as something with the capacity to mine lots of data to recognize patterns and then to make recommendations….It’s similar in many cultures to the council of elders,” Clarke noted. “So why do we have council of elders? Because two heads are better than one. Because crowdsourcing produces better outcomes.”
Creating AI-First Gov: New Policy Initiatives
Following the first panel, participants had the opportunity to practice using AI to address critical public policy issues. In an exercise facilitated by D^3, Lakhani, Rem Koning, Associate Professor at Harvard Business School, and Jin Paik, Co-Founder and Managing Partner of Altruistic, participants leveraged chatbots like ChatGPT 4.0 to analyze city-level data from Austin, Texas.
Through structured phases and guided prompts, participants developed a policy proposal using ChatGPT that addressed Austin’s most pressing needs. This method was then applied to policy development for cities in emerging economies, demonstrating the adaptability, as well as the potential limitations, of AI as a policy tool.
Adapting AI for Emerging Economies: Insights from Global Innovators at GEM25
In the second panel, AI Innovations in Emerging Economies, Zain Verjee, Founder and CEO of theZVG and former CNN Anchor, moderated a discussion on how innovators are adapting AI for emerging economies. Panelists shared their strategies for addressing market frictions and overcoming regulatory, financial, and infrastructural hurdles.
One central theme was the potential of AI to tackle foundational challenges in areas like education, healthcare, and agriculture. Panelists noted that in many emerging economies, the focus is on using AI to bridge gaps in essential services.
HE Ahmed Alsuwaiyan, Governor of the Digital Government Authority of Saudi Arabia, stressed that AI should not be seen as an objective in itself, but rather as a method to drive specific outcomes. Alsuwaiyan also emphasized the need to break down job roles into specific tasks when considering AI’s impact on the workforce. “Whenever we think about AI or generative AI replacing jobs, we have to look into the job taxonomy. We have to break down the job profile. Start with the tasks and workflows, but don’t start with the full job profile,” he explained.
The panelists echoed the need to align AI solutions with local contexts. Rebecca Enonchong, Founder & CEO of AppsTech, noted that Africa is not a homogeneous market and that innovations must account for the continent’s cultural and legal diversity. She also reflected on how innovators have developed solutions, despite these variable and challenging environments. “We’re entrepreneurs first, so we find solutions and those solutions are to look outwards for a market. We’re not going to let ourselves fall behind on AI because our government is not there for us,” Enonchong said. “We are finding solutions and we are building solutions from the continent, for the continent, and beyond, despite the hostile environment.”
Jalal Charaf, Chief Digital Officer for University Mohammed VI Polytechnic (UM6P) highlighted how Moroccan universities are leveraging AI to transform education and agriculture, with custom-built avatars and AI tutors contextualized for different learning environments. “The key word here is context — this technology allows us to localize and customize at scale," he shared.
Collaboration emerged as another major factor in driving AI innovation. Paula Bellizia, Vice President at Amazon Web Services and President of Latin America, emphasized the power of partnerships across governments, the private sector, and academia in creating impactful solutions. “When all parts of society come together with the purpose of solving challenges, you start to see the seeds of what can be the right purpose moving forward,” she remarked.
In addressing regulatory hurdles, Gustav Praekelt, Founder of turn.io, focused on the importance of data integration. He called for private-public collaboration within countries and data sharing across government entities, in addition to cross-border efforts. According to Praekelt, the answer is to have the end user in mind and end-user integration across service delivery nodes is critical.
GEM25 Closes with a Celebration of Global Development and AI Innovation
GEM25’s five interactive Incubation Rooms provided a platform for participants to address critical public challenges, brainstorm multi-sectoral solutions, and engage with a powerful network of experts and leaders. This year’s Incubation Rooms explored the following topics:
- Competition and Industrial Policy in the Era of AI
- AI and Education
- Regulation and Governance of the Digital Commons
- Firms and Labor in the Age of AI
- AI and Global Health Equity
Participants were encouraged to develop collaborations within these sessions for submission to the GEM Incubation Fund, which provides seed funding for research proposals seeking to leverage AI to solve key development challenges.
Looking forward: AI and Global Empowerment
GEM25 concluded with a closing reception, wrapping up a full day of lively discussion, collaboration, and innovation. The conference was the annual culmination of CID’s Road to GEM programming, which highlights leaders in development shaping the future with AI through CID events, podcasts, and articles.
A special thank you to the GEM25 speakers and facilitators of our GEM Incubation Rooms, to the guests who traveled to Boston and engaged with passion and commitment, to the GEM25 Scientific Committee, and to our staff, dining, and custodial teams for ensuring the event ran smoothly. Check out the conversations on LinkedIn, BlueSky, Instagram, and X using #GEM25.
Image Credits
Matt Teuten
Read Next Post
View All Blog Posts
