*OpenAI has introduced “Images in ChatGPT,” allowing users to generate images directly within ChatGPT using GPT-4o.
This feature is available across Free, Plus, Pro, and Team subscription tiers, with the free tier having a usage limit similar to DALL-E’s. Fans of DALL-E can still access it via a custom GPT, The Verge reports. Research lead Gabriel Goh describes this model as a major leap forward, leveraging GPT-4o’s “omnimodal” foundation, which integrates text, image, audio, and video generation. A key improvement is the “binding” of attributes with objects, ensuring that AI maintains proper relationships between them. Unlike previous models, which struggled with multiple elements, this version can accurately handle 15 to 20 objects in an image.
“If I go to draw an image, I do so with the limitation of my own skill… but also with all of the knowledge of the world that I’ve built up,” ChatGPT multimodal product lead Jackie Shannon explained.
“The model brings world knowledge to the equation, so when you ask for an image of Newton’s prism experiment, you don’t have to explain what that is to get an image back,” Shannon added.
Text rendering has also improved, reducing errors in AI-generated text on images.
Scroll to continue reading
“While we certainly have room to improve on latency…the quality of these images, the capability, the world knowledge, really makes up for the additional seconds that they’ll spend waiting,” Shannon said.
Although image generation requires more time, OpenAI finds the improvement in quality to be a worthwhile tradeoff.
“While we certainly have room to improve on latency…the quality of these images, the capability, the world knowledge, really makes up for the additional seconds that they’ll spend waiting,” Shannon said.
The system includes safeguards against misuse, preventing watermark removal, explicit deepfakes, and CSAM. While images lack visible AI watermarks, they contain C2PA metadata to indicate OpenAI as the creator.
“Ultimately, no system is perfect for this type of thing, but we’re continuously improving our safeguards and we think of this as a starting point,” Shannon said. “One thing that’s true about all of the images generated from ChatGPT is that the user owns them and are free to use them within the bounds of our usage policies as they would like.”
READ MORE FROM EURWEB.COM: Trump Calls the Smithsonian’s National Museum of African American History ‘Divisive, Race-centered, Corrosive’ – Threatens Funding
Sign up for our Free daily newsletter HERE.
