Earlier this year, OpenAI introduced DALL-E, an AI-based tool that generates images from a textual description. If you’ve joined OpenAI’s waiting list, get ready: the company has selected the lucky users who get the chance to test out DALL-E. One million people have been selected, and if you get to be among them, the developers have shared some details to help you out.
In the announcement, OpenAI writes that they will invite the selected candidates from the waitlist over the coming weeks. If you are selected, you will get 50 free credits during your first month of use, and 15 free credits every month to follow. You can use one credit in two different ways. One, for a single original DALL-E prompt generation, each one returning four images. Or two, an edit or a variation prompt, which returns three images.
This is the first phase of the beta, and if you wish, you can buy additional DALL-E credits in 115-credit increments for $15 on top of your free monthly credits. Each purchase like this is around 460 images worth, depending on how you use the credits. “One credit is applied each time a prompt is entered and a user hits ‘generate’ or ‘variations’, OpenAI explains. You can even use the generator for commercial projects.
Is it safe?
With AI technology, safety is always one of the biggest concerns. OpenAI writes that they had worked with researchers, artists, developers, and other users before launching the beta. The company claims that it has “taken steps to improve [its] safety systems,” and this includes:
Curbing misuse: To minimize the risk of DALL-E being misused to create deceptive content, we reject image uploads containing realistic faces and attempts to create the likeness of public figures, including celebrities and prominent political figures. We also used advanced techniques to prevent photorealistic generations of real individuals’ faces.
Preventing harmful images: We’ve made our content filters more accurate so that they are more effective at blocking images that violate our content policy — which does not allow users to generate violent, adult, or political content, among other categories — while still allowing creative expression. We also limited DALL-E’s exposure to these concepts by removing the most explicit content from its training data.
Reducing bias: We implemented a new technique so that DALL·E generates images of people that more accurately reflect the diversity of the world’s population. This technique is applied at the system level when DALL-E is given a prompt about an individual that does not specify race or gender, like “CEO.”
Monitoring: We will continue to have automated and human monitoring systems to help guard against misuse.
Artists of all kinds are welcome to join the waiting list, as well as journalists. So naturally, I applied as well and still keeping my fingers crossed that I’ll get invited. You can still join the waitlist for app access. And in case you get invited or need inspiration, check out this trippy music video or these non-existent cameras created with DALL-E.
[via PopPhoto; image credit: OpenAI]