OpenAI has introduced new functionalities to enhance its AI-powered image generator DALL-E, specifically within its ChatGPT interface. The latest additions, unveiled on Wednesday, enable users to modify images generated by DALL-E directly within ChatGPT and provide style prompts for creative inspiration. These enhancements aim to make the tool more accessible and user-friendly, especially for casual users. Notably, OpenAI also made ChatGPT accessible to all users, regardless of whether they have an OpenAI account, on Tuesday.
The announcement was shared by OpenAI's official account on X (previously known as Twitter). The in-line editing feature is now available on the web client, as well as the Android and iOS apps of ChatGPT. Users with a premium subscription to OpenAI can visit ChatGPT, and enter a prompt to generate an image, and if they are unsatisfied with the result, they can open the image to access an edit icon.
The edit icon, located before the download option, allows users to select specific areas on the image for editing. After selecting the desired areas, users provide a new prompt describing the changes they want. The AI then regenerates the image based on the input changes. Notably, Microsoft's Copilot Designer also introduced a similar in-line editing feature recently.
Non-Graphic Designers May Face Some Challenges
For individuals without a background in graphic design, navigating DALL-E's native platform can be challenging due to its complexity, requiring detailed prompts and style weights for desired outcomes. OpenAI addressed this issue by integrating the AI image generator with ChatGPT after the launch of DALL-E 3 in September 2023. This integration aimed to simplify the process for users, making it easier to fine-tune the images they create using the editing tool.
Additionally, OpenAI implemented a feature in ChatGPT where style prompts are displayed above the text field to inspire users. Styles like Woodcut, Close-up, Low angle, Artificial lighting, and more are presented, with hovering over each style revealing an example. Users can select their desired style and provide a simple prompt to generate an image in the chosen style.