OpenAI has integrated an upgraded AI image generator directly into ChatGPT
OpenAI has integrated an upgraded AI image generator directly into ChatGPT and text-to-video model, Sora (called “Images in ChatGPT”), effectively replacing its original text-to-image model, DALL-E, which marks the first major update to ChatGPT’s image-generation capabilities in over a year. - ‘Images in ChatGPT’ is powered by OpenAI’s reasoning model, GPT-4o, so it has better world knowledge and contextual understanding, meaning itcan create/modify more accurate and detailed images and photos. - Although it will “think” for longer before creating an image, it will understand contextual prompts and it will also produce coherent and accurate text within images, something AI image generators struggle with. - OpenAI has openly confirmed that to power the new image generator, it trained it on “publicly available data” and proprietary data from its partnerships with companies like Shutterstock. - - Although many AI companies, including OpenAI, usually prefer to keep what data they used to train their AI models close to their chests, OpenAI has chosen to reveal its training data sources because it wants to respect “artists’ rights,” and has confirmed it has “policies in place” that prevent it from generating images that “directly mimic any living artists’ work,”as they continue to battle several ongoing copyright lawsuits.