Midjourney plans to rollout an updated web tool that would enable users to edit any images uploaded from the web using Midjourney's generative AI.
Updated tool
The updated tool, according to Midjourney CEO David Holtz can be rolled out "early next week," and can also be used to "repaint" the color and details of objects in images through retexturing.

This is a hot-button topic these days on editing existing images with AI. At Meta, they're wrestling over how to label images that are edited with AI versus generated from scratch with an AI model, while Google has released powerful AI features that give no visual indication that images are AI-modified.

In 2022, Midjourney said it would add the IPTC's Digital Source Type property, a technical standard embedding metadata in images that signify them to have AI-generative origins. Yet, of the largest AI platforms, only Elong does not use C2PA, which embeds metadata tracing an image's full provenance.

The enhanced image tool - which, Holz wrote in a post on the official Discord server of Midjourney, will first be limited to "a subset of the current community" with enhanced human moderation and "new, more advanced AI moderators" to try to prevent abuse.

"Honestly, we're not sure how to precisely limit the rollout of this feature," he said. Midjourney's taking a community poll and using that input to figure out who sees the feature first.

There are risks to launching these kinds of editing tools without proper controls in place. They could make it easy to commit copyright infringement on a scale that dwarfs previous cases or be used to push deepfakes that deceive more people.

Deepfakes are sweeping social media like fire, making it challenging to discern reality from disinformation. Artificial intelligence-generated images of destruction and human suffering following Hurricane Helene recently flooded the web.

Data from Clarity, a deepfake detection firm, shows that there have been 900% more deepfakes created and published this year compared to this time last year. This is pretty frightening for obvious reasons. The latest YouGov poll of Americans revealed that 85% believed the misuse of deepfakes should be a reason for concern regarding the spread of falsehoods online.

With no federal law in the U.S. that criminalizes deepfakes, over 10 states have passed statutes outlawing AI-assisted impersonation. A California law awaiting enactment will be the first to allow judges to issue removal orders or even impose monetary sanctions on poster holders of deepfakes.

And Midjourney hasn't exactly been a shining example of responsible AI deployment. For one, it's being sued over its alleged use of copyrighted content to train its generative AI models. But in recent months, the platform has taken steps to limit the spread of deepfakes-heck, filters for political figures leading up to the U.S. presidential election.