Midjourney plans to let anyone on the web edit photos with AI

Midjourney plans to release an advanced web tool that will allow users to edit any image uploaded to the web using Midjourney’s generative AI.

The improved tool, which Midjourney CEO David Holtz said will be released “early next week,” will also allow users to repost objects in photos to “repaint” their colors and details according to captions.

Editing existing images with AI has become a hot topic recently. Platforms like Meta have struggled with how to label images edited with AI tools versus those produced from scratch with an AI model, while companies like Google have released powerful AI features that don’t provide visual reference images replaced with AI.

Last year, Midjourney committed to using the IPTC’s Digital Source Type structure, a technology standard that embeds metadata in images that show they are AI-generated. However, the company is one of the few major AI platforms that has yet to adopt C2PA, a metadata technology that tracks the full evolution of an image, including the hardware and software used to create it.

In a post on Midjourney’s official Discord server, Holz said the improved photo tool will be limited to “a small group of the current community” to begin with, with increased human moderation and “new, more advanced AI moderators” to try to prevent it. torture.

“Honestly, we’re not sure how to precisely limit the use of this feature,” he said. Midjourney is asking for public feedback via a survey that it will use to determine which users get access first.

There is a risk of releasing these types of editing tools without adequate safeguards in place. They may organize large-scale copyright infringement, or encourage the spread of misleading deepfakes.

Deepfakes spread like wildfire on social media, and make it difficult to distinguish fact from disinformation. Recently, fake AI images of destruction and human suffering have flooded the web after Hurricane Helene.

According to data from Clarity, a deepfake detection company, 900% more deepfakes have been created and published this year compared to the same period last year. It’s alarming, understandably so. A recent poll from YouGov found that 85% of Americans are concerned about misleading deepfakes circulating online.

While there is no federal law criminalizing deepfakes in the US, more than 10 states have enacted laws against AI-assisted impersonation. The California law – which is currently on hold – would be the first to give judges the power to order deepfakes posters to take them down or face fines.

Midjourney was never a good example of responsible AI deployment. (It’s been sued for allegedly using copyrighted content to train its artificial intelligence models, for example.) But in recent months, the platform has taken steps to limit the spread of deepfakes, including filters for political figures leading up to the US president. election.