SI Agent News
DevelopingClaudeOpenAI

ChatGPT adds GIF creation and persistent photo memory

OpenAI's new Images 2.6 update allows users to turn prompts into looping animations and save specific photos for reuse in future conversations.

50% fasterimage generation speed increase

The short version

Why it matters

These features move ChatGPT from a one-off image generator toward a more persistent creative tool. By allowing for animations and saved references, the AI can maintain visual consistency across multiple projects.

Faster and more consistent visuals

OpenAI recently released Images 2.6, a new model designed to improve the speed and reliability of visual generation. The update is now available to all subscription plans worldwide across web, mobile, and desktop platforms. It also integrates into the Codex coding tool and the API for developers.

One of the primary technical improvements is that images now generate 50% faster than they did previously. This means users can see their results in less time while maintaining the quality of the output. The company also claims that the model handles edits with greater consistency during the generation process.

This consistency is particularly important for maintaining specific details like faces, clothing, or product designs. When a user asks to change a background or a minor detail, the model is better at keeping the rest of the subject the same. This reduces the frustration of having a character's appearance shift randomly between different iterations.

Moving beyond still images

The update introduces a new capability for creating looping GIFs from either text prompts or uploaded photos. Users can describe the specific movement they want to see, or they can provide a still image and ask the AI to animate it. Once the animation is created, the playback speed can be adjusted by the user.

While the tool allows for the creation and sharing of these animations, there are currently limitations on editing them. As of now, ChatGPT cannot edit an animated GIF that a user has uploaded to the platform. The system is currently rolling out, so the feature may not be visible to every user immediately.

The rollout has already been spotted in some regions, such as Germany, where templates for anime reaction GIFs appeared on the website. OpenAI has not yet published specific limits on how many GIFs can be created on different subscription plans. Users can save or share their finished GIFs directly from the chat interface.

Persistent photo memory

A second major feature allows users to save personal photos and reference them in future conversations using a naming system. Users can upload a picture, assign it a name, and then call it up later using an @mention. For example, a user could save a photo of a dog named Pepper and later ask for a watercolor portrait of @Pepper.

This system removes the need for users to re-upload the same image every time they want to generate a new variation. Additionally, ChatGPT may offer to save a photo automatically if a user provides a detailed description of a specific subject. Once saved, these photos stay stored within the user's account for easy access across different chat sessions.

There is a distinction between these saved references and standard generated images. Generated images are automatically saved under a specific Images section and are only deleted when the conversation that created them is removed. However, the help page does not currently specify where users can manage or delete their saved photo references.

Safety and data privacy

The ability to save photos and create animations raises potential concerns regarding the depiction of real people. Combining a saved reference photo with animation makes it easier to place individuals into scenes that never actually occurred. OpenAI has not yet stated how it plans to guard against the misuse of GIFs that feature real people.

To help identify AI-generated content, images created with ChatGPT include both Content Credentials and an invisible SynthID watermark. These are tools used to signal that a piece of media was produced by an artificial intelligence. However, the company has not listed GIFs separately in its current documentation regarding these provenance signals.

Users are encouraged to only save photos of children or other people with their explicit consent. Regarding data usage, any photos uploaded to the chat are part of the conversation history. Users can control whether OpenAI uses these uploads for training purposes by adjusting their settings in the Data Controls section of their account.

What to watch

Try it

Upload a photo of a pet, give it a name like @Buddy, and see if you can generate a new style of portrait using that name in a new chat.

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by Notebookcheck · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news