AI for Equity Must Transfer Power, Not Only Tools

A major artificial-intelligence commitment can sound like equity before anyone has received a service, learned a new capability or gained a meaningful decision right. The announced amount is visible. The institutional arrangements that determine who defines the problem, owns the data, selects the model, evaluates the result and carries the recurring cost are much less visible.

Money matters. Many schools, training institutions, clinics, farms and ministries work under severe resource constraints. Appropriate investment can extend knowledge, strengthen professional judgement and make useful services available in languages and places commercial markets neglect. Refusing every externally funded technology would protect neither agency nor opportunity.

The governing question is therefore not whether philanthropy should support AI. It is whether support produces equitable power. A community has not gained equity merely because a tool reaches it. Equity requires authority to shape the purpose, protect people, evaluate benefit, correct failure, retain capability and continue serving when the funder or provider changes direction.

On 15 September, the Associated Press reported that the Gates Foundation is pledging US$1 billion over two years to expand what it describes as equitable artificial intelligence. The foundation's 2026 Goalkeepers report frames the central choice starkly: AI may narrow opportunity gaps or intensify them, depending upon decisions made now. Associated Press 2026 Goalkeepers report

According to the announcement, roughly US$400 million is intended for education, US$400 million…