OpenAI announced the release of a major update to its image-generating AI tool DALL-E.
DALL-E 3 is the first upgrade to the tool since DALL-E 2 debuted in April 2022 and provides significantly improved outputs.
The tool is in research preview now and will be generally available to ChatGPT Plus and Enterprise users in October.
With the update, users can utilize ChatGPT text prompts to create and refine pictures and have the resulting images delivered right in the chat interface.
OpenAI has trained the AI using over 12 billion parameters and unlike many competing image generators the connection to ChatGPT makes it much simpler to create requests.
DALL-E reportedly provides an endless variety of outputs even with the same prompt.
OpenAI CEO Sam Altman has made safety and security a major tenet of OpenAI products and has spoken often of this need including in front of a US Senate committee earlier this year.
As part of the DALL-E 3 release, the company has announced that any image requests that reference living artists will be refused.
So users can request a painting using Monet but not Banksy.
But this topic is far from over as the discussion of the usage of original materials is still much in question.
A recent lawsuit was filed against OpenAI by a number of well-known authors over the use of copyrighted materials that were used to train the generative language model for ChatCPT.
Authors John Grisham, George R.R Martin, Michael Connelly, and many others have joined together in a class action lawsuit that mirrors other similar lawsuits against Meta Platforms and Stability over the use of copyrighted material.
Leave a Reply