Website: https://openai.com/dall-e-2/ , https://labs.openai.com/
Brief Description: DALL-E 2 is an AI system that can create realistic images and art from a description in natural language. You can use DALL-E 2 to generate images that combine concepts, attributes, and styles, such as “a classroom in the shape of an aircraft.” DALL-E 2 is an improved version of DALL-E, with 4x greater resolution and enhanced consistency. You can also use DALL-E 2 to edit existing images or expand them beyond the original canvas. Explainer Video: DALL·E 2 Explained
User Policy: This is an interesting article on copyright issues related to AI-generated products. As the article mentions, the U.S. Copyright Office’s “copyright law only protects works that are made by a human being, not those of a monkey, an elephant, or an AI model. If a human didn’t author the work, the Copyright Office won’t register the copyright (and you can’t sue someone if you can’t register your copyright). In the eyes of the Copyright Office, the public is free to reproduce, publish, or sell DALL-E 2-generated masterpiece, no strings attached… As it stands, you can pretty much do anything commercially that you want with your DALL-E 2 image.” Moreover, according to the OpenAI, “you can use Content for any purpose, including commercial purposes such as sale or publication if you comply with these. Here Content means collectively the input you provide and the output you receive in OpenAI platforms, such as ChatGPT, DALL-E 2, etc. This is the link to their terms of use- https://openai.com/terms-of-use
Cost: The initial account registration is free. It offers 50 initial monthly credits that you can use for image creation. Once you have consumed all your credits, each next month, you will then have 15 free credits refilled instead of 50 credits. Otherwise, you can pay simultaneously to purchase even more credits to avoid worrying about running out. You can use these free credits differently depending upon the type of picture resolution output you desire. The price will vary depending on the resolution type.
Paper: https://arxiv.org/abs/2204.06125
Reference: Ramesh, A., Dhariwal, P., Nichol, A., Chu, C., & Chen, M. (2022). Hierarchical Text-Conditional Image Generation with CLIP Latents. ArXiv. /abs/2204.06125