OpenAI has introduced the latest iteration of its text-to-image tool, DALL-E 3, employing ChatGPT to help generating prompts. This new feature will enhance the user experience by allowing them to request images with more accurate requests, as prompts will be adjusted through interactions with ChatGPT itself.
With DALL-E 3, users can simply input their image request and refine it through conversations with ChatGPT. This, through natural language interactions, can enhance even brief prompts, providing more descriptive guidance to the model. The overall process will result in highly detailed and accurate images.
Notably, the upgraded software brings about advancements in generating higher-quality images, particularly when prompted with more extended sentences. Moreover, it addresses historical challenges faced by image-generating models, such as accurately representing human hands and dealing with text.
In terms of safety measures, DALL-E 3 has new mechanisms to minimize algorithmic bias and improve content safety. It rejects requests for images in the style of living artists or depicting public figures. Additionally, OpenAI allows artists to opt out of having their artwork used to train future versions of text-to-image models. This has been done to address copyright concerns.
This release intensifies the competition in the generative AI domain, especially in image synthesis. Other companies like Midjourney and Stability AI are continuously refining their models, putting pressure on OpenAI and its competitive edge.
Initially, this new version will be available to premium ChatGPT users in October, followed by availability to research labs and API customers.