OpenAI is launching wider availability of its latest text-to-image generator. On Thursday, the company will give ChatGPT Plus and Enterprise customers access to the new DALL-E 3 model that works within the ChatGPT app. OpenAI says it has prepared a security mitigation stack for the model that prepares it for an expanded release.
DALL-E 3 was first announced last month, and OpenAI showed how it improved on the previous DALL-E 2 by allowing users to take advantage of ChatGPT to type longer, more visually descriptive prompts to feed the image generator. DALL-E 3 was added to Bing Chat and Bing Image Generator, making Microsoft’s platform the first to introduce broader public access to the model, even before ChatGPT.
Guardrails advertised to mitigate harmful images have not always worked, with users generating images of the World Trade Center while SpongeBob SquarePants and other characters fly planes into the buildings. Even after Microsoft blocked certain prompts, other simple fixes produced similar results.
Text-to-image generators like Midjourney, Stable Diffusion, and earlier versions of DALL-E have had their fair share of controversy. Technology has produced copyrighted image materials, non-consensual nudity, ethnic changes of subjects, and photorealistic misrepresentations of public figures.
OpenAI promises that it has taken much broader steps this time and is provide a website that shows research conducted on DALL-E 3. The company says it will “limit the model’s likelihood of generating content in the style of living artists, images of public figures and improve demographic representation in the generated images.” OpenAI also has an internal “provenance classifier” tool that it says is capable of detecting with 99 percent accuracy whether an image was generated by DALL-E 3.