Key Insights from GEM25 Panel at Harvard
At the recent GEM25 event hosted by Harvard’s Center for International Development, policy and technology experts from emerging economies discussed integrating Artificial Intelligence (AI) within their national strategies. Panelists included Teresa Clarke, Chairman and CEO of Africa.com; Nanjira Sambuli, Nonresident Scholar at Carnegie Endowment for International Peace; Dr. Shikoh Gitau, CEO of Qhala; Umar Saif, Founder and CEO of aiSight.ai; moderated by Han Sheng Chia from the Center for Global Development.
Examining Technology Adoption: Electricity vs. AI
Moderator Han Sheng Chia opened with a cautionary perspective, referencing research that challenges the intuitive assumption that electricity directly improves household income. Citing studies from Kenya and J-PAL, he highlighted how electricity connections, surprisingly, did not significantly raise incomes or educational outcomes as widely expected. Chia used this analogy to prompt reflection on whether the adoption of AI similarly translates directly into meaningful developmental outcomes.
Mobile Internet as AI’s Immediate Predecessor
Dr. Shikoh Gitau countered the analogy with electricity, emphasizing clear economic benefits from mobile internet adoption observed in Africa. Gitau highlighted how internet connectivity significantly boosted job opportunities and household incomes. She specifically pointed to AI’s potential to improve healthcare access, mentioning an AI-powered pathology diagnostic tool that greatly reduces cancer diagnosis times in regions with severe shortages of medical specialists.
Gitau underscored, “Africa has 17% of the world’s population but only 1% of its medical professionals,” emphasizing AI’s critical role in addressing this gap.
Survey Highlights: Optimism and Concerns about AI in Africa
Teresa Clarke shared insights from an Africa.com survey, highlighting strong optimism (78%) among respondents from Nigeria, Kenya, and South Africa regarding AI’s potential positive impacts, especially in education, healthcare, and agriculture. However, concerns were raised about job displacement, ethical issues, and lack of transparency. Interestingly, 55% of African respondents worried about excessive government regulation, significantly higher than comparable views in the U.S.
Respondents emphasized the necessity of culturally sensitive AI applications. One participant stated, “Solutions should consider cultural, linguistic, and socioeconomic factors to effectively address unique African challenges.”
Electricity and Infrastructure: Critical Foundations for AI
Panelists stressed the critical role of infrastructure—especially electricity—in realizing AI’s benefits. Teresa Clarke noted AI’s enormous power demands, comparing the energy consumption of U.S. AI infrastructure to Japan’s total national electricity usage. The panel agreed that substantial investments in reliable electricity and digital infrastructure are essential for meaningful AI adoption in emerging economies.
AI for Business: Opportunities and Realities
Umar Saif introduced practical skepticism about AI applications in developing regions, highlighting issues of high costs and limited data availability. However, he described how his company, aiSight.ai, successfully uses satellite imagery, crowdsourced local data, and mobility analytics to offer actionable insights to global corporations in emerging markets.
“These companies lack basic market data,” Saif explained, detailing how aiSight.ai helps firms like Coca-Cola optimize distribution, pricing, and inventory strategies in previously underserved areas.
AI as a Strategic Partner, Not Just a Tool
Clarke emphasized AI’s role as a strategic partner rather than merely a technical tool. She explained how AI changes the dynamic between technology and decision-makers in both businesses and governments. Clarke described AI as resembling a “council of elders,” providing data-driven insights tailored specifically to organizational objectives, such as cost reduction or service enhancement in telecommunications.
Balancing Data with Human Insight
Nanjira Sambuli highlighted the indispensable role of human context and cultural understanding alongside AI capabilities. She illustrated this with the example of Kenyan fintech companies initially attempting to restrict nighttime lending to curb gambling, only to learn from local insights that market women also depended on such loans for early morning trading. Sambuli argued that human judgment remains crucial for appropriately interpreting AI-generated insights.
Representation and Data Equity
The panel stressed the importance of diverse representation in AI development to avoid embedded biases. Clarke referenced Joy Buolamwini’s research on facial recognition biases, highlighting the necessity for diverse coding teams to create equitable AI systems that accurately reflect diverse populations.
Localization of Data Assets
Finally, panelists emphasized the importance of localizing data infrastructure. Clarke stressed that local data centers are vital for competitive global access, underscoring the strategic significance of local control over digital assets.
Overall, GEM25 panelists highlighted AI’s vast potential but emphasized that successful implementation requires careful, localized, and culturally aware approaches.
Watch the full panel discussion below.
