[ad_1]
Mid-flight It plans to release an upgraded web tool that will allow users to edit any images downloaded from the web using Midjourney’s generative AI.
The upgraded tool, which Midjourney CEO David Holtz said will be released “early next week,” will also allow users to recompose objects in photos to “repaint” their colors and details according to the captions.
Editing existing photos using AI has become a hot topic recently. Platforms like Meta have grappled with how to classify images edited using AI tools versus those created from scratch using an AI model, while companies like Google have issued… Powerful AI features Which images do not give a visual indication that they have been modified by artificial intelligence.
Last year, Midjourney committed to using the IPTC’s Digital Source Type attribute, a technology standard that embeds metadata into images that indicate they were created by artificial intelligence. However, the company is one of the few major AI platforms that has not adopted C2PA, a metadata technology that tracks the entire provenance of an image, including the hardware and software used to create it.
In a post on The official Discord server for MidjourneyThe upgraded photo tool will be limited to “a subset of the existing community” to start, with increased human moderation as well as “new and more advanced AI moderators” to try to prevent abuse, Holz said.
“Honestly, we’re not sure how to precisely restrict the deployment of this feature,” he said. Midjourney is soliciting community feedback through a survey that it will use to determine which users get access first.
There are risks in launching these types of editing tools without adequate safeguards. They can facilitate widespread copyright infringement, or encourage the spread of misleading deepfakes.
Deepfakes are spreading like wildfire across social media, making it even more difficult to distinguish between truth and misinformation. More recently, fake generative AI images of devastation and human suffering have spread It flooded the Internet In the aftermath of Hurricane Helen.
According to data from Clarity, a deepfake detection company, 900% more deepfakes were created and disseminated this year than in the same time frame last year. It causes alarm, and that’s understandable. A recent YouGov poll found this 85% of Americans were worried About misleading deepfakes spreading across the Internet.
In the absence of a law criminalizing deepfakes at the federal level in the United States, more than a dozen states have enacted laws against artificial intelligence-assisted impersonation. California’s law – currently stalled – would be the first to enable judges to order deepfake posters to be removed or potentially face financial penalties.
Midjourney has not been a shining example of responsible AI deployment. (It’s being sued for allegedly using copyrighted content to train generative AI models, for example.) But in recent months, the platform has taken steps to curb the spread of deepfakes, including filters for political figures leading up to the US presidential election. election.
[ad_2]