OpenAI Introduces Direct Editing of AI Images Within ChatGPT

ChatGPT’s image generation capabilities have received an exciting upgrade. OpenAI recently announced on X that users of GPT-4 powered ChatGPT can now interact with generated images by clicking on them to activate a paintbrush-style tool. This tool allows users to highlight specific sections of the image for adjustments or enhancements.

In a demonstration video, a user creates an image of a dog wearing a conical party hat. Upon clicking the image, the user enters an editing view, selects the paintbrush tool, and highlights the dog's head. They then type the prompt “add bows,” resulting in a new version of the image that maintains the original while adding red bows in the designated area.

Several users noted that these features were quietly rolled out earlier this week. A media outlet tested the new functionality on a personal account and confirmed that it works as intended.

This feature is powered by OpenAI’s DALL-E 3 model, which was introduced in September 2023. The integration of DALL-E 3 into ChatGPT enhances its capabilities, providing users with increased utility and control over image generation.

In addition to the new painting tool, OpenAI has introduced a style inspiration feature for DALL-E 3, allowing users to choose from pre-set styles like woodcut, close-up, low angle, artificial lighting, and hand-drawn.

These advancements occur amidst ongoing legal challenges facing OpenAI and increasing criticism from some artists and users regarding AI-generated art. Concerns center around models like DALL-E 3 being trained on copyrighted images without permission or compensation.