Key Takeaways
- The World Bank advocates for “small AI” – lightweight models that run on mobile phones with low connectivity and in local languages – to bring AI benefits to developing economies.
- Paschal Donohoe, the Bank’s chief knowledge officer, argues that small AI could compress development timelines from centuries to decades.
- Small AI relies on narrowly‑trained models for specific tasks, such as soil‑analysis advice for Indian farmers in the AgriConnect programme.
- Successful deployment will require collaboration among the World Bank, national governments, and the private sector to overcome infrastructure gaps.
- IMF Managing Director Kristalina Georgieva stresses that electricity, digital infrastructure, skills training, and regulatory safeguards are preconditions for AI to be widely adopted and “as democratic as the cellphone.”
- Georgieva warns that up to 40 % of global jobs may be augmented or eliminated by AI, with young entrants to the labour market most at risk.
- Under President Ajay Banga, the World Bank has mobilised roughly $120 bn of private capital in the past year, using guarantees and other financial tools to spur investment in emerging markets despite shrinking aid budgets in donor countries.
World Bank Pushes for “Small AI” in Developing Nations
The World Bank is urging low‑ and middle‑income countries to embrace a new class of artificial intelligence it calls “small AI.” Unlike the massive, data‑hungry models powering Silicon Valley’s consumer gadgets, small AI consists of specialised models trained on narrower datasets for a defined purpose. Paschal Donohoe, the Bank’s chief knowledge officer, explained that these models can operate on a basic mobile phone even when connectivity is poor and can deliver information in the user’s local language. “We argue that economic development in the countries that we serve and partner with that could have taken centuries to make happen, the potential is there for that economic development now to happen within decades,” Donohoe said.
What Small AI Looks Like on the Ground
To illustrate the concept, Donohoe cited the World Bank’s AgriConnect initiative in India, which links farmers with agronomic advice. “There’s not going to be enough scientists and agricultural experts to do that for every small farm in India,” he noted. “And this is where what we describe as ‘small AI’ comes in. Because small AI is AI that would be available on your phone, that can help deliver against those information needs.” In practice, a farmer could snap a photo of a soil sample, upload it via a low‑bandwidth connection, and receive a recommendation in Hindi or a regional dialect—all without needing a high‑end smartphone or constant internet access.
Collaboration Needed to Scale Small AI
Donohoe stressed that realizing the promise of small AI will not happen through the World Bank alone. He called for a tripartite partnership involving governments, private‑sector tech firms, and the Bank itself. Governments would need to create enabling policies and invest in basic digital infrastructure, while companies could adapt their AI tools to run on modest hardware. The Bank, for its part, can facilitate financing, share knowledge, and help de‑risk investments that might otherwise deter private actors from entering nascent markets.
IMF Chief Highlights Preconditions for AI Success
Speaking at the same World Bank‑IMF annual meetings in Bangkok, IMF Managing Director Kristalina Georgieva echoed the need for foundational readiness. “You need access to electricity. You need access to digital infrastructure. People have to be geared to adoption. And the government should be able to give some safety, protect against bad actors so that adoption can work,” she said. Georgieva added that if these basics are in place, AI adoption can become “very democratic. As democratic as the cellphone.” Her remarks underscored that without reliable power and connectivity, even the most lightweight AI models will remain out of reach for many communities.
Job Displacement Concerns Loom Large
Georgieva also warned of the labour‑market implications of AI, estimating that roughly 40 % of jobs worldwide could be either augmented or eliminated by the technology over time. She singled out young people entering the workforce as particularly vulnerable. “Normally they would get into these routine, repetitive jobs and then go up the ranks. But if they’re gone, I don’t think any young person wants to be a barista in a coffee shop as a professional future,” she remarked. The warning serves as a reminder that AI’s benefits must be paired with policies that support reskilling, social safety nets, and the creation of new, higher‑value occupations.
World Bank’s Annual Development Report Flags Infrastructure Gaps
The Bank’s own World Development Report, released in August, highlighted the stark realities that hinder AI adoption in many poorer regions. It noted that “almost a third of rural schools in sub‑Saharan Africa lacked electricity and more than two‑thirds did not have a consistent internet connection.” Such statistics illustrate why the push for small AI—designed to function intermittently and on low‑power devices—is both timely and necessary. The report implicitly argues that AI strategies must be tailored to the constraints on the ground rather than assuming the high‑bandwidth, always‑online environments typical of wealthy economies.
Private Capital Mobilisation as a Development Lever
Under President Ajay Banga, the World Bank has sharpened its focus on attracting private‑sector finance to emerging markets. Donohoe revealed that the Bank had mobilised roughly $120 bn (≈ £90 bn) of private capital in the past year through guarantees, blended finance, and other instruments. This figure is especially notable given that traditional aid donors such as the United States and the United Kingdom have been trimming their overseas assistance budgets. “We want to highlight the link between private capital mobilisation and jobs and economic development,” Donohoe said, suggesting that leveraging private money can accelerate the deployment of technologies like small AI while simultaneously creating employment opportunities.
Conclusion: A Pragmatic Path Forward for AI in the Global South
The World Bank’s advocacy for small AI reflects a pragmatic shift: instead of chasing the cutting‑edge models that demand massive data centres and constant high‑speed connections, the focus is on lean, purpose‑built tools that can work where infrastructure is limited. Success, however, hinges on meeting basic preconditions—electricity, connectivity, skills, and regulation—as emphasized by Georgieva. It also requires addressing the disruptive potential of AI on jobs, especially for youth, and coupling technology rollout with robust reskilling programmes. Finally, the Bank’s recent success in pulling private capital into developing economies offers a financing pathway that could make small AI a reality for millions of farmers, entrepreneurs, and students who currently sit on the wrong side of the digital divide. If these elements align, the promise of compressing centuries‑long development trajectories into decades may move from aspiration to achievable policy.
https://www.theguardian.com/technology/2026/oct/18/world-bank-artificial-intelligence-developing-economies