Bill Gates has cast artificial intelligence as a choice with unusually high stakes: it could become “the greatest equalizer” or the “worst source of injustice”. The Gates Foundation is putting $1 billion behind its preferred outcome, committing the money over the next two years to expand access to AI in health care, agriculture and education, with its work most focused on Africa and South Asia.
The pledge was announced alongside the foundation’s annual 2026 Goalkeepers report, titled Make This Matter: AI, Equity, and the Choice We Can't Delay. The report argues that governments, technology companies and philanthropies have the next 12 to 18 months to act before the direction of AI development hardens around the needs of wealthy countries and customers.
Gates has described that prospect as a market failure. The most capable systems are being built first for the wealthiest users, he said, while the foundation’s report warned that AI will prioritise profit over need without deliberate intervention. The foundation’s wager is that funding tools for underserved communities early can change that path.
About $400 million of the commitment is intended for health care and a further roughly $400 million for education. Agriculture will receive about $100 million. An additional $100 million is earmarked for AI training data sets in languages spoken by underserved communities.
The language allocation addresses one of the clearest gaps in current systems. AI speech-recognition tools make errors more than 60% of the time in Yoruba, compared with less than 6% in English, according to the foundation. Such disparities can limit the usefulness of systems intended for communities whose languages have received little attention from the companies building them.
Health care is another area where the foundation sees an opening. In sub-Saharan Africa, there is roughly one doctor for every 2,000 people, compared with fewer than 200 people per doctor in high-income countries. The foundation is seeking to put AI tools in the hands of health workers facing those constraints, rather than reserving them for well-resourced hospitals and markets.
OpenAI is contributing $50 million to a pilot programme placing AI tools in Rwandan clinics. The work follows a January commitment by the foundation and OpenAI of $50 million for Horizon 1000, a multi-year rollout of digital health tools across 1,000 clinics by 2028, beginning in Rwanda. Gates visited Rwanda in July to review the project.
Other technology groups have been assigned more specific roles. Google.org and Microsoft’s AI for Good Lab are backing an open call for proposals on language infrastructure. Anthropic is working with the foundation on efforts to accelerate vaccine research and broaden AI knowledge of locally grown crops.
The announcement adds an AI dimension to a larger plan for the foundation’s remaining years. In May 2025, it said it would spend more than $200 billion over two decades and close in 2045, while doubling its rate of giving. Gates said in June 2025 that most of the funding would go towards Africa. The foundation also announced a $1.4 billion agricultural commitment for Africa and Asia in November 2025.
The new pledge arrives as AI systems spread globally while their training data and early deployments remain concentrated in English-language and wealthy markets. The foundation has named the central problem plainly, though important details remain unresolved. It has not set out how it will measure success or track equity outcomes, which health conditions, crops or educational areas will take priority, or how much funding will reach the poorest regions rather than middle-income countries with stronger AI infrastructure.
The announcement also leaves open whether the resulting tools will be open-source, proprietary or a mixture of both. It has not described enforcement or accountability arrangements for ensuring that corporate partners give equity priority in their work. Gates has also acknowledged the darker uses of the technology, saying AI will empower people to carry out fraud, disinformation, surveillance, cyberattacks and bioterrorism. The foundation’s challenge is therefore broader than distributing software: it is trying to influence who AI serves while the industry’s commercial habits are still being formed.
Comments (0)