Meta’s New AI Image Editor Is Using People’s Instagram Photos Without Permission
Alexandr Wang, chief AI officer at Meta.
Credit: Bloomberg/Getty Images
Meta’s latest AI endeavour isn’t off to a great start, as it’s angered Instagram users over manipulating images from public profiles without user permission.
Muse Image is the hottest new AI model from Meta’s Superintelligence Labs, the division tasked not just with making AI as smart as humans, but much smarter. And yet its latest release is a new slop generator that can quickly edit images, create new ones, or mash up existing images into something else entirely. It’s not particularly distinct from any of the other options out there; indeed, Meta’s own testing shows it falling behind OpenAI’s Images 2.0, though beating Google’s Nano Banana 2 in editing and multi-image prompts.
The big difference with this one, though, is it’s tied directly into Instagram, WhatsApp, and the Meta AI app. And unless you opt out, it can use your images, too.
If you’re an Instagram user with a public-facing profile, you will need to explicitly opt out of allowing Muse Image to access your photos. Otherwise, all someone has to do is tag you in a prompt, and your photos and images are fair game for remixing and generative editing.
The reason Meta doesn’t seem too concerned with the pushback (yet) is that this tool isn’t really for Instagram users—it’s for advertisers. As CNBC explains, Muse Image is also powering Meta’s Advantage Plus service, which uses AI to enable brands to easily develop creative adverts and modifications to those adverts; think automated, advanced A/B testing.
This is the kind of AI tool that analysts have been encouraging Meta to develop for a while: one that enhances its core business model of selling adverts, using its enormous user base as a testing bed for each ad. It has the potential to be a strong earner for Meta, creating a revenue stream that could help offset at least some of the tens of billions it’s spent on AI development and infrastructure over the past few years.
But in making Muse Image public, and likely garnering a lot more training data for the trouble, Meta has opened itself up to the kind of criticism previously levied at xAI’s Grok over unsanctioned use of people’s images for AI manipulation. Although the profiles are public and therefore tacitly consent to anyone seeing those images, it’s unlikely every account owner with a public profile will be happy for their images to be used by AI, especially since many of those photos were uploaded years before AI photo manipulation was even a thing.
Still, Meta is ploughing ahead. It has announced plans to release Muse Video in the future, offering the same kind of manipulation and advertising content-generation capabilities, but with video instead of images.
