ChatGPT’s Image Generation Feature Receives Significant Upgrade

On Tuesday, during a live stream, Sam Altman, the CEO of OpenAI, announced a major update to ChatGPT's image generation capabilities, marking the first significant upgrade in over a year.

ChatGPT now utilizes OpenAI's GPT-4o model to create and modify images and photos natively. Previously, GPT-4o supported only text generation and editing, but with this upgrade, it expands its functionality to visual content.

According to Altman, the new image generation feature is available today for users subscribed to the company's $200 per month Pro plan, both in ChatGPT and Sora, OpenAI's video generation product. OpenAI plans to roll out the image generation capabilities soon to Plus and free users of ChatGPT, as well as to developers who access the company’s API services.

The GPT-4o model, now equipped with image output functionality, processes requests slightly longer than its predecessor, DALL-E 3. OpenAI describes the result as more accurate and detailed images. This model allows users to edit existing images, making changes that can include the transformation of certain elements or inpainting specific details in both the foreground and background.

OpenAI has not disclosed the specific image datasets used to enhance the new image generation functionalities. Many companies in the generative AI space view their training data as a competitive edge and typically keep these details confidential. Additionally, revealing specifics about training data could lead to potential intellectual property lawsuits.

For creators who want to exclude their works from OpenAI's training datasets, the company provides an opt-out form. Furthermore, OpenAI is committed to respecting requests to prevent its web-scraping bots from collecting data from various websites, including images.

This upgrade to ChatGPT's image generation feature comes shortly after Google introduced its own experimental image generation capabilities with the launch of Gemini 2.0 Flash. The release generated significant interest on social media, although it also attracted criticism for lacking protective measures, which allowed users to remove watermarks and create images featuring copyrighted characters.

ChatGPT, image, upgrade, OpenAI, technology