Correcting the Map Must Reach the AI Curriculum

A learner opens a world map and sees Africa reduced beside northern landmasses. Another asks an artificial-intelligence system to compare African cities, universities, churches or cultures and receives a confident answer shaped by uneven data. A governing body studies a digital dashboard whose categories make some communities visible and others statistically absent. These are not separate problems. Each representation selects, scales and organises reality before a person begins to reason with it.

No flat map can preserve every property of a globe, and no curriculum, dataset or model can contain the whole world. The ethical question is not whether representation involves selection. It is whether the purpose and distortion are understood, whether excluded people can challenge the result, and whether institutions mistake a useful model for neutral truth.

Correcting a map is therefore more than changing an image on a classroom wall. It is an invitation to govern the complete chain through which knowledge is produced and acted upon: purpose, source, classification, representation, platform, interpretation, assessment, decision, correction and public accountability.

On 4 September, the United Nations General Assembly adopted the Togo-sponsored Correct the Map resolution by 164 votes to one, with six abstentions. Backed by the African Union, it encourages governments and institutions to move from the familiar Mercator world map towards an equal-area representation such as Equal Earth in education, media and public…