Midjourney plans to release an upgraded web tool that will allow users to edit images uploaded from the web using Midjourney's generated AI.
The upgraded tool, which Midjourney CEO David Holtz said will be released “early next week,” will also allow users to retexture objects in images and “redraw” colors and details according to captions. You will be able to do it.
Editing existing images with AI has become a hot topic these days. Platforms like Meta struggle with how to label images edited with AI tools versus images generated from scratch with AI models, while companies like Google We've released a powerful AI feature that doesn't visually indicate that it's changed.
Last year, Midjourney committed to using IPTC's Digital Source Type property, a technology standard that embeds metadata in images to indicate they were generated by AI. However, the company is one of the few major AI platforms that has not adopted C2PA, a metadata technology that tracks the full provenance of images, including the equipment and software used to create them.
In a post on Midjourney's official Discord server, Holtz said the launch of the upgraded image tools will be limited to a “subset of the current community,” will include increased human moderation, and will include “new, more advanced AI. moderator” will try to prevent it. abuse.
“Honestly, I don't know how exactly to limit the deployment of this feature,” he said. Midjourney is seeking feedback from the community through a survey that we will use to determine which users will get access first.
Releasing these types of editing tools without proper safeguards creates risks. These could facilitate large-scale piracy or facilitate the spread of misleading deepfakes.
Deepfakes are spreading like wildfire across social media, making it difficult to separate truth from disinformation. Most recently, Hurricane Helen flooded the web with fake AI-generated images depicting destruction and human suffering.
According to data from deepfake detection firm Clarity, more than 900% more deepfakes were created and published this year compared to the same period last year. Understandably, it's causing alarm. A recent YouGov poll found that 85% of Americans are concerned about misleading deepfakes being spread online.
The United States has no federal law criminalizing deepfakes, and more than 10 states have enacted laws banning AI-based impersonation. California's law, which is currently stalled, would be the first to give judges the power to order deepfake posters to take them down. Failure to do so may result in fines.
Midjourney is not a shining example of responsible AI implementation. (For example, the platform has been sued for allegedly using copyrighted content to train generative AI models.) But in recent months, the platform has been using filters for politicians ahead of the U.S. presidential election. , has taken steps to limit the spread of deepfakes. election.