Midjourney is going to release a better web tool that will let people use Midjourney’s generative AI to edit any web images that they share
Midjourney CEO David Holtz said the updated tool will be available “early next week.” It will also let users “repaint” the colors and details of items in pictures by changing their texture.
These days, using AI to change existing photos has become a hot subject. Platforms like Meta have been trying to figure out how to distinguish between pictures edited with AI tools and images created from scratch with an AI model. At the same time, companies like Google have released powerful AI features that don’t make it obvious when AI has changed images.
Midjourney promised to use the IPTC’s Digital Source Type property last year. This technical standard adds information to pictures that show AI made them. However, the company is one of the few big AI platforms that haven’t adopted C2PA, a metadata technology that shows where a picture came from and what hardware and software were used to make it.
Holz said in a post on Midjourney’s official Discord server that the improved image tool will only be available to a “subset of the current community” initially. To stop abuse, there will be more human moderation and “new, more advanced AI moderators.”
“To be honest, we don’t know how to limit the deployment of this feature precisely,” he said. Midjourney uses a poll to get feedback from the community, which it will then use to decide which users get entry first.
There are risks when these changing tools are made public without the right safety measures. They could make it easier for huge amounts of copyright violations to happen or help spread fake deepfakes.
