OpenAI’s DALL-E has received a big boost that brings new editing chops to the AI platform. The first version of this AI tool was announced back in January 2021. Three years later, the third generation of the software DALL-E 3 was released natively to ChatGPT, ChatGPT Plus and ChatGPT Enterprise customers in October. The new update for the third generation of DALL-E has now added image editing tools when using it within ChatGPT.
It is available on both web and mobile. The second new feature added to DALL-E is preset style suggestions that help in enhancing image creation. It is quite similar to the AI-generated wallpaper prompts that you get on Android. OpenAI even sent out a post on X, informing users about the new update. You can now edit DALL-E images in ChatGPT across the web, iOS and Android.
A newly added ‘select’ button enables users to highlight specific sections of images that users want to edit. One can also change the design of the objects in an image or add new ones. The company has also added several usability features including the undo and redo buttons. Users can even adjust the aspect ratio of the image while also accessing new style suggestions.
The new style prompts in the DALL-E-3 service help the users with a starting point. It can provide them with different styles like woodcut, gothic, synth-wave, and hand-drawn with examples. Choosing one of the styles will add it to your prompt and then it’s onto the user to figure out what else they want in the image.
Given the recent additions, it is likely that OpenAI is currently working on making its image-generating software tool more user-friendly. Its integration with ChatGPT takes more burden off user’s shoulders when it comes to writing a perfect description of an image prompt. One could simply tell the software what you want with it doing the rest. The latest update is more of an extension of that flexibility where a user can now tweak and refine their image results directly within the ChatGPT interface.
OpenAI has also recently been putting in some work to address the “trust” concern surrounding AI image creation. Earlier this year, the image-generating software received visible watermarks and metadata to segregate fake images from the real ones. While this move is a good start, both of these indicators can be easily removed using third-party software. OpenAI will now be looking to resolve this problem to address the pressing concern.
Comments
0 comment