OpenAI ChatGPT4 subscribers now have premium access to the DALL-E 3 image generator within the app.
OpenAI has announced a broader release of its new text-to-image creation tool. Starting this Thursday, ChatGPT Plus and Enterprise users will get to experience the new DALL-E 3 feature within the ChatGPT application. OpenAI has implemented safety measures to ensure the model’s readiness for a wider rollout.
The introduction of DALL-E 3 came just last month, highlighting its advancements over the earlier DALL-E 2 version. This new model lets users work with ChatGPT to craft more extended and vivid descriptions, which can then be transformed into images. Notably, DALL-E 3 debuted on Microsoft’s Bing Chat and Bing Image Generator platforms, giving the public a taste of its capabilities even before ChatGPT did.
However, not all the safety measures to prevent inappropriate imagery have been foolproof. There were instances where users created inappropriate images related to the World Trade Center. Despite Microsoft’s efforts to block certain requests, users found alternative methods to obtain similar inappropriate images.
Text-to-image tools like Midjourney, Stable Diffusion, and previous versions of DALL-E have faced challenges and controversies. Concerns have arisen over the technology creating copyrighted content, inappropriate images, altering the ethnicity of individuals, and making misleading images of well-known personalities.
Addressing these issues, OpenAI assures that they’ve taken comprehensive measures for DALL-E 3. They have even set up a website showcasing the in-depth research behind the model. The company is committed to reducing the chances of the model creating images that mimic living artists’ styles, images of celebrities, and ensuring better demographic diversity in the images it produces. Furthermore, OpenAI has developed an in-house tool that boasts a 99% accuracy rate in identifying images generated by DALL-E 3.