OpenAI has introducedDALL-E 3, and it takes text-to-image to new heights.
As you might know, DALL-E was one of the pioneers of generative art. It’s also a sister app to ChatGPT. As it’s often said, teamwork makes the dream work, and the pair have finally pooled their powers to bring more precision to your prompts and image output. Natively integrated with ChatGPT, DALL-E 3 can better translate text inputs into vivid artwork.
Simply toss ChatGPT anything from a simple sentence to a detailed paragraph, and DALL-E 3 will whip up dreamy and detailed images that complement—and build upon—your vision. That’s because the chatbot is able to interpret even vague descriptions and come up with vivid stories.
The trick here is that ChatGPT does a swell job understanding what you’re saying and serves up tailored prompts to DALL-E 3, so your ideas pop on the screen. If an image isn’t quite what you envisioned, a few words to ChatGPT can swiftly refine it.
In a video, OpenAI demonstrates how a parent got the language model to dream up designs for a “super-duper sunflower hedgehog” named Larry.
According to OpenAI, the upgraded text-to-image tool yields visibly superior results to its predecessor DALL-E 2. And like DALL-E 2, artwork produced on DALL-E 3 is entirely yours to use without restrictions.
With an emphasis on safety, DALL-E 3 is further enhanced with safeguards to prevent the generation of abusive or adult content. Additionally, it blocks queries mentioning public figures or the styles of living artists, ensuring responsible usage. For creators, OpenAI is exploring the possibility of opting out of having their work trained on its tools, per Reuters.
DALL-E 3 is currently in its research preview stage and will roll out to ChatGPT Plus and Enterprise customers via the API in October, with availability in Labs later this fall.