Multilingual artificial intelligence often reinforces bias

A new study from Johns Hopkins finds that multilingual AI privileges dominant languages, deepening divides rather than democratizing access to information