A marketing team wants to automatically create brand-new promotional product images from short text descriptions like 'a red sneaker on a mountain trail at sunset.' They have no existing images to start from — the visuals must be generated entirely from the text prompts. Which type of AI model should they use?