Languages the machines skip

Global in scope·A current case, last reviewed September 2026

Machine translation is strong between data-rich languages and weak or absent for thousands of others, so speakers of low-resource languages get worse search, worse moderation and thinner access to the web's knowledge. The models learned from what was digitised, and digitisation had already chosen favourites.

The concepts this raises

Each of these gives a class a different way into the same case.

PowerKnowledge and Society
The capacity to shape what counts as knowledge, who is heard, and what gets funded, published or taught. Power is not only censorship; setting an agenda, controlling a platform or being granted authority all change the knowledge that a community ends up with.
CultureThe Knower's Lens
The shared practices, language and inherited assumptions of a community of knowers. Culture supplies the questions that seem worth asking and the answers that seem obvious, so its influence is hardest to see from inside it.

Where it sits in the knowledge framework

Ethics
The ethical questions internal to the knowledge itself: what may be investigated and how, what obligations the knowledge places on those who hold it, and the ethical constraints the discipline imposes on its own practitioners.

Areas of knowledge

Themes

Cases that raise a concept this one raises, but in a different area of knowledge or a different part of the world. That contrast is the point: it is what shows a student the concept travels.