- calendar_today August 10, 2025
OpenAI launched “Images in ChatGPT,” which brings revolutionary image generation capabilities straight to the ChatGPT interface. The new GPT-4o model enables this breakthrough feature, which allows users to generate images during their conversations, thereby advancing AI-driven content creation substantially.
The new feature “Images in ChatGPT” unlocks advanced image generation for users across every subscription tier, including Plus, Pro, Team, and the free version. OpenAI spokesperson Taya Christianson stated that free tier users will experience similar usage limits to DALL-E 3 at approximately three images per day, but acknowledged that these restrictions may change with demand. OpenAI ensures DALL-E enthusiasts will have access through a specially created GPT.
OpenAI’s research lead Gabriel Goh called GPT-4o an “omnimodal” foundation that works with multiple data types such as text, images, audio, and video. The model now features improved binding functionality which solves a prevalent issue found in AI image creation. The GPT-4o model reliably maintains relationships between objects and attributes in images with 15 to 20 objects and avoids color or shape errors, which earlier models could not manage.
The system demonstrates an exceptional advancement through its enhanced ability to render text. AI-generated images usually display text that appears distorted and meaningless. Goh described the development as an iterative process that required several months to perfect. The team recognizes perfect text rendering as unattainable with small text elements, but they have reached a consistency standard to make text in images dependably usable.
The system uses an autoregressive architecture instead of the diffusion models, which most image generators rely on. The sequential image generation method from left to right and top to bottom mimics text generation and is believed to enhance text rendering and binding performance.
OpenAI presented its system’s numerous uses in a briefing through demonstrations of scientific diagram generation with exact labels, consistent character multi-panel comic creation with proper dialogue, and poster designing, which included correct text. Demonstrations included practical applications like producing stickers with transparent backgrounds, together with restaurant menus and logos.
Jackie Shannon, who leads multimodal products at ChatGPT, highlighted how the system utilizes world knowledge. When she starts drawing an image, she uses her personal artistic abilities but also incorporates everything she knows about the world. The model incorporates world knowledge, which allows you to request an image of Newton’s prism experiment without needing to provide any additional background information.
OpenAI believes that the improved quality and expanded capabilities of image generation make the slightly longer wait time worthwhile. The quality of these images and their capabilities, combined with extensive world knowledge, compensate for the extra waiting time despite our need for latency improvements.
Safeguards and User Ownership: Ensuring Responsible AI Image Generation
OpenAI demonstrated a commitment to addressing misuse possibilities through its adoption of strong protective measures. The system implements safeguards that block watermark removal and sexual deepfake generation while rejecting CSAM content requests. Generated images will bear the standard C2PA metadata, which identifies them as OpenAI creations since visual watermarks are not present. The company operates internal systems to verify images internally.
Shannon acknowledged that while no system is flawless for these purposes, they are dedicated to enhancing their protective measures with this functioning as an initial step. Users retain ownership of all images created by ChatGPT and can utilize them according to our usage policies as they see fit.
OpenAI’s “Images in ChatGPT” feature not only improves its main product but extends AI-driven creative potential by enabling visual expression through its conversational platform. OpenAI demonstrates its dedication to user satisfaction through this new feature while actively addressing the potential dangers of advanced AI image generation capabilities. The pursuit of better binding capabilities, along with text rendering enhancements and security safeguards, demonstrates a commitment to building a powerful yet accountable tool.
