AI learns to create images from text descriptions by destroying data

Early last year, artificial intelligence company OpenAI unveiled software with the surprising ability to create accurate images from text captions – even obscure inventions such as "an armchair in the shape of an avocado". The company has now released a new AI model that is smaller but capable of producing even better results. Last year's program – called DALL-E – was a large, 12-billion-parameter AI model that was trained on a huge set of images with associated captions.