Digital art created on DALL-E. Image: OpenAI
With more than 1.5m active users of DALL-E now creating more than 2m images a day, OpenAI hopes making it public will further improve the tech.
OpenAI has removed the waitlist for DALL-E to give wider access to the text-to-image model and further improve the AI technology.
In an announcement on its website yesterday (28 September), OpenAI said that people wishing to avail of the tool can now sign up and starting using the service straight away.
DALL-E – an AI that can create images from simple word descriptions – was first developed last year. But a newly unveiled DALL-E 2 in April generates “more realistic and accurate images with four times greater resolution”.
OpenAI said the tool now has more than 1.5m active users who create more than 2m images a day. This includes artists, creative directors, authors, architects and regular users – with more than 100,000 people sharing their creations in DALL-E’s Discord community.
“Responsibly scaling a system as powerful and complex as DALL-E – while learning about all the creative ways it can be used and misused – has required an iterative deployment approach,” the company wrote on its website.
“Since we first previewed the DALL-E research to users in April, users have helped us discover new uses for DALL-E as a powerful creative tool. Artists, in particular, have provided important input on DALL-E’s features.”
OpenAI expanded access to its beta over the summer. Feedback from users so far has helped it add new features to DALL-E such as Outpainting, which lets users continue an image beyond its original borders and create images of any size, and Collections, which lets users “create in new ways” and expedite their creative processes.
“Learning from real-world use has allowed us to improve our safety systems, making wider availability possible today,” the company added.
Ahead of the waitlist removal, OpenAI said it developed new “detection and response techniques” and made its filters “more robust” at rejecting attempts to create images that contain adult or violent content – or anything that violates its content policy.
OpenAI is now working on a DALL-E API with testing currently underway – soon to be offered to developers and businesses to build apps on the system. “We can’t wait to see what users from around the world create with DALL-E,” the company wrote.
Last week, OpenAI started letting users upload and edit people’s faces on DALL-E.
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.
