According to Mari-Sanna Paukkeri of Utopia Analytics, an AI model to identify hate speech can be built for any language in merely two weeks, if provided a definition of hate speech in a particular culture.