AI Image Generator DALL-E is now available in limited open beta

DALL-E

DALL E 2, OpenAI’s artificial intelligence-powered image generator that creates realistic images and art from a natural language description, is now available in limited beta form to the public. Until now, the powerful AI tool was only available to a select group of users after it was announced in April.

The San Francisco-based company announced on Wednesday that it will soon send invitations to one million people on its waiting list for the latest version in the coming weeks.

For inexperienced users, DALL·E 2 takes a phrase or series of words as input and outputs professional-looking paintings or hyper-realistic photographs. The DALL E 2 also includes the following features:

Edit: This allows users to perform realistic and context-aware edits to images they generate using DALL·E or images they upload using natural language description.

variations: This can take an image generated by DALL·E or an image uploaded by a user and make different variations of it inspired by the original.

My collection: Allows users to store generations in the DALL·E platform.

“DALL·E allows users to create quickly and easily, and artists and creative professionals use DALL·E to inspire and accelerate their creative processes. We’ve already seen people use DALL·E to create music videos for young cancer patients, create magazine covers and bring new concepts to life,” the company wrote in a blog post.

Each DALL·E user receives 50 free credits for the first month of use and 15 free credits every following month. Each credit represents four images based on one original DALL·E prompt, or three images if the user provides an edit or variation prompt.

Users can also purchase additional DALL·E credits in increments of 115 credits (460 images) for $15 on top of their free monthly credits. Each time a prompt is entered and a user presses “generate” or “variations”, one credit is awarded.

Furthermore, artists in need of financial assistance can also apply for subsidized access, OpenAI says.

The beta version also allows users to get full usage rights to commercialize the images they create with DALL·E, including the right to reprint, sell and merchandize. This includes images they generated during the survey preview.

OpenAI says it has already attracted interest from users who plan to use DALL·E images for commercial projects, such as illustrations for children’s books, art for newsletters, concept art and characters for games, mood boards for design advice and storyboards for films.

OpenAI adds that it has worked with researchers, artists, developers and other users to ensure its technology is not misused for malicious purposes or to create misinterpretations and deep forgeries.

The company says it is taking steps to reduce the creation of misleading content by refusing to upload images with realistic faces. It has also made its content filters more accurate, preventing users from generating violent, mature, or political content, among other categories. In addition, it has implemented a new technique in DALL·E that creates images of people that more accurately reflect the diversity of the world’s population and do not specify race or gender.

OpenAI emphasizes that the DALL-E 2 will continue to have automated and human surveillance systems to prevent abuse.

“Expanding access is an important part of our responsible deployment of AI systems, as it allows us to learn more about real-world use and continue to iterate on our security systems,” OpenAI added. . “We continue to explore how AI systems, such as DALL-E, can reflect biases in the training data and different ways we can address them.”

Leave a Comment