Some big changes have been made to OpenAI’s DALL-E that give the AI tool new editing skills. In January 2021, the first version of this AI tool was made public. After three years, in October of this year, the third generation of the software DALL-E 3 was made available to ChatGPT, ChatGPT Plus, and ChatGPT Enterprise users. With the new update for the third version of DALL-E, you can now edit pictures when you use it in ChatGPT.
You can get it on both the web and your phone. The second new thing that DALL-E has is preset style ideas that can help you make better images. It resembles the AI-generated wallpaper suggestions that Android gives you in a lot of ways. OpenAI even wrote a post on X to let people know about the new version. You can now change DALL-E pictures in ChatGPT on the web, iOS, and Android.
The “select” button that was just added lets users pick out parts of pictures that they want to change. You can also add new items or change the way the ones that are already there look. The company has also added some features that make it easier to use, such as the undo and restart buttons. Users can even change the image’s aspect ratio and get new style ideas at the same time.
The new style prompts in the DALL-E-3 service give people a place to start. It can show them samples of arts and styles like gothic, hand-drawn, woodcut, and synth-wave. Once you pick a style, it will be added to your message. The user can then choose what else they want in the picture.
With the new features, OpenAI may be working on making its software tool for making images easier for people to use right now. Because it works with ChatGPT, users don’t have to worry as much about writing a perfect description of a picture prompt. You could just tell the software what you want it to do, and it would do it. With the most recent update, users can now tweak and improve their picture results right in the ChatGPT interface, which makes it even more flexible.
OpenAI has also been working hard lately to handle the “trust” issue that comes up when AI makes pictures. The software that makes pictures got watermarks and metadata earlier this year that can be seen to tell the difference between real and fake images. This is a good start, but both of these signs are easy to get rid of with software from a different company. OpenAI will now try to fix this issue to take care of the most important matter.