OpenAI, the artificial intelligence research company, has announced the launch of Dall-E 3, a new version of its text-to-image tool that uses its popular AI chatbot ChatGPT to help users generate prompts. Dall-E 3 will be available to ChatGPT Plus and Enterprise customers in October via the API, the company said in a blog post.
What is Dall-E 3 and how does it work?
Dall-E 3 is an improved version of Dall-E, a generative AI visual art platform that converts text prompts to images. Dall-E was first released in January 2021 and has since amazed the internet with its ability to create realistic and surreal images from simple or complex descriptions. For example, Dall-E can produce images of “an armchair in the shape of an avocado” or “a painting of a capybara sitting in a field at sunrise” based on the text input.
The new version of Dall-E, however, has a new feature that integrates with ChatGPT, OpenAI’s widely used AI chatbot that can generate coherent and engaging conversations from natural language input. By using ChatGPT, users do not have to come up with their own detailed prompt to guide Dall-E 3; they can just ask ChatGPT to come up with a prompt, and the chatbot will write out a paragraph for Dall-E 3 to follow. Users can also tweak the prompt through conversations with ChatGPT.
For example, in a demo shown by OpenAI, a user asked ChatGPT to help them create a logo for a ramen restaurant in the mountains. ChatGPT then wrote a longer prompt, such as “a logo for a ramen restaurant in the mountains that features a bowl of ramen with steam rising from it and a mountain peak in the background”. Dall-E 3 then generated four options for the user to choose from.
This connection with the chatbot, OpenAI said, allows more people to create AI art because they do not have to be very good at coming up with a prompt. Dall-E 3 can also understand context better and produce more accurate and detailed images than the previous versions.
What are the challenges and opportunities of text-to-image AI tools?
OpenAI’s race to create accurate text-to-image AI tools has several competitors, including Alibaba’s Tongyi Wanxiang, Midjourney and Stability AI, who continue to refine their image-generating models. These tools have various applications in fields such as education, entertainment, design, marketing and journalism. They can also inspire creativity and innovation among users who want to explore the possibilities of AI art.
However, there are also several challenges and concerns around AI-generated images. One of them is the ethical and legal issue of intellectual property rights. A Washington D.C. court in August ruled that a work of art created by AI without any human input could not be copyrighted under U.S. law. This raises questions about who owns and benefits from the AI-generated images and whether they infringe on the rights of human artists.
Another challenge is the potential misuse and abuse of text-to-image AI tools for creating harmful or hateful content. OpenAI said it has worked hard to implement robust safety measures to prevent Dall-E 3 from generating violent, adult or hateful content. It has also added mitigations to decline requests that ask for images of a public figure by name or those that ask for images in the style of a living artist. Moreover, OpenAI said it will allow creators to opt out of using some or all of their work used to train future text-to-image tools.
OpenAI said it hopes that Dall-E 3 will enable more people to express their ideas and imagination through AI art. It also said it will continue to monitor and improve the tool’s performance and safety as it releases it to more users in the coming months.