OpenAI unveils DALL-E 3, allows artists to opt out of training

Image credits: Bryce Durbin/TechCrunch

OpenAI today unveiled an upgraded version of its text-to-image tool, DALL-E, that uses ChatGPT — OpenAI’s viral chatbot — to take some of the pain out of prompting.

Most of today’s sophisticated AI-powered image creation tools take prompts — image descriptions — and turn them into artwork in a range of styles, from photorealistic to fantasy. But crafting the right vector can be a challenge, so much so that “agile engineering” has become a real profession.

OpenAI’s new tool, DALL-E 3, uses ChatGPT to help fill out claims. Via ChatGPT, subscribers to OpenAI’s premium ChatGPT plans, ChatGPT Plus and ChatGPT Enterprise, can write a request for an image and refine it through conversations with the chatbot – receiving the results directly within the chat app.

ChatGPT will take a short prompt of no more than a few words and make it more descriptive, providing more direction to the DALL-E 3 model.

ChatGPT integration isn’t the only thing new in DALL-E 3. DALL-E 3 also creates higher-quality images that reflect prompts more accurately, OpenAI says — especially when dealing with longer prompts. It handles content that has historically hampered image generation models better, such as text and human hands.

Image created by DALL-E 3.

Furthermore, DALL-E 3 has new mechanisms to reduce algorithmic bias and improve safety, or so says OpenAI. For example, DALL-E 3 will reject requests requesting a portrait in the style of living artists or depictions of public figures. Artists can now choose not to use some — or all — of their artwork to train future generations of OpenAI models to convert text to image. (OpenAI, along with some of its competitors, is facing a lawsuit for allegedly using artists’ copyrighted works to train its generative AI image models.)

See also  Artificial intelligence is mastery of language. Should we trust what he says?

The launch of DALL-E 3 comes as the race for generative AI heats up, especially in the field of image synthesis. Competitors like Midjourney and Stability AI continue to improve their image-generating models, putting pressure on OpenAI to stay afloat.

OpenAI plans to roll out DALL-E 3 to premium ChatGPT users in October, followed by research labs and its API customers. The company hasn’t said when — or if — it plans to release a free web tool, as it did with DALL-E 2 and the original DALL-E model.

Leave a Reply

Your email address will not be published. Required fields are marked *