Recently, ChatGPT, along with various AI chatbots, has integrated image generation capabilities. Many users have likely encountered this feature, particularly as its recent advancements have gained significant attention.

While discussions surrounding the ethical implications of AI-generated art and its various inaccuracies and peculiarities are important, it’s worth noting a few challenges with ChatGPT’s image generation. Users often find it cumbersome to track their creations, as each image remains linked solely to the conversation in which it was generated. This limitation makes organizing and retrieving images a tedious task unless one manually catalogs them elsewhere. OpenAI, however, has addressed this issue with an innovative solution.

Unveiling the ChatGPT Image Library

On a recent Tuesday, OpenAI introduced the Image Library, a new component of ChatGPT. This feature automatically archives all images produced using ChatGPT’s 4o model throughout your conversations, eliminating the need to sift through previous discussions in search of a specific image. Think of it as similar to your smartphone’s default photo gallery, where any snapshots taken or saved converge into a single, organized location.

This Tweet is currently unavailable. It might be loading or has been removed.

All tiers of ChatGPT users—Free, Plus, and Pro—can now access the Image Library directly from both the mobile app and web platform. To find it, simply look for the “Library” option in the left-hand side panel, positioned underneath “Explore GPTs.” (While the web application should update automatically, users may need to refresh their ChatGPT app for the new feature to appear.)

Upon entering the Library, users will encounter their prior image creations. It’s important to note, however, that it might not display all images generated. Some users have reported noticing only a handful of their recent creations, despite having produced more images historically. OpenAI clarifies that legacy DALL-E images will not populate this section; only creations made with the 4o model will be included. Nevertheless, OpenAI has indicated that it is in the process of backfilling the library, meaning additional older images could become visible over the coming days.

Selecting an image from the library brings up a variety of options: Edit, Select, Save, and Share. The “Edit” function allows for initiating a new chat featuring the selected image, enabling users to instruct ChatGPT on desired modifications. The “Select” option permits users to emphasize a specific section of the image for editing, ensuring that only the chosen area is revised. The “Save” feature facilitates storing images to the user’s device library, while “Share” provides the capability to send the image to various platforms.

While image generation via ChatGPT may not be a primary function for every user, this addition certainly streamlines the experience for those who do utilize it frequently. However, one notable drawback remains: the absence of a “delete” feature. To remove an image from the ChatGPT library, users must navigate to the conversation in which the image was first generated and delete it from there. Although the Library allows access to the original thread containing the image, deleting an individual image or the associated thread must be accomplished separately by locating the conversation title in the sidebar. This process complicates the user’s experience, and it would be beneficial if OpenAI considered integrating a straightforward “delete” option for each image in the library in the future.