After people tirelessly searched for ways to get access to OpenAI’s text-to-image generator, DALL-E 2, the company has finally announced its release without the waitlist. OpenAI announced the tool DALL-E for the first time in January 2021, creating a lot of buzz around AI text-to-image generators. Then people had to signup for a waitlist in order to access the tool. DALL-E 2 attracted a lot of eyes and everyone was practically dying to get access. The company recently opened up the beta version of DALL-E 2 for 1 million users, and today it announced it is available for everyone.
DALL-E 2 now has more than 1.5 million users and they are now creating more than 2 million images every day. The company also mentioned that releasing this tool has helped them in discovering new ways that it can be used and feedback from especially artists has helped them release many features like the outpainting feature.
Although an API is not available as of now, the company also announced that it is working on one, and will be available soon. This will help businesses and developers build apps using the system.
Although when DALL-E was released it impressed many people, the company released the tool very cautiously and drove the attention to other systems like Stable Diffusion and Midjourney. DALL-E is a little late to the party as these tools are already available to a wide audience. On top of that, Stable Diffusion is also being used by people to create explicit images, meaning lesser restrictions. But OpenAI did this in order to prevent users from misusing the tool.
“Learning from real-world use has allowed us to improve our safety systems, making wider availability possible today. In the past months, we’ve made our filters more robust at rejecting attempts to generate sexual, violent and other content that violates our content policy and built new detection and response techniques to stop misuse.”
The company said in a blog post, “Learning from real-world use has allowed us to improve our safety systems, making wider availability possible today. In the past months, we’ve made our filters more robust at rejecting attempts to generate sexual, violent and other content that violates our content policy and built new detection and response techniques to stop misuse.”
AI systems like DALL-E are trained on images available all over the internet which is why image generations that can create harm in society are practically only a few clicks away. As these AI tools are developing, there are also many legal as well as ethical questions that surround them. It is unclear whether AI-generated creations are stolen or not. Recently, Getty Images and a few other similar services banned the material because of concerns that it might violate copyright. It seems AI text-to-image generators still have some issues that they need to address until the world welcomes them with open arms.
Featured Image Credits: OpenAI