AI image oddness stems from underrepresented languages and poor translation

Evidence

Journalist Jason Koebler traced the oddness of the images to “prompts that are written in Hindi, Urdu, and Vietnamese, which are underrepresented in large language model training data.” He continued: “Other bizarre outcomes arise because people are using Google Translate and speech-to-text to say a prompt aloud in Hindi, translate it to English, and are using that often poorly-translated prompt to generate the images.” — W. David Marx, Blank Space, 698

Categorization

• QUESTION: This suggests the ‘bizarre’ aesthetic of contemporary internet imagery is a language-bias artifact of the AI stack — worth chasing how this shapes what ‘weird’ content even looks like.