Meta on Tuesday unveiled Muse Image, its new AI image generator built by Meta Superintelligence Labs, the company’s dedicated AI unit. The feature, which was internally code-named Mango, is now available for free through the Meta AI app, as well as on Instagram Stories and WhatsApp. Unfortunately, the new model is already causing controversy.

What exactly can you do with Muse? It sounds like the use cases are similar to most other AI image generators — you’ll be able to create plenty of goofy, cartoonish images, for instance. If you’re short on inspiration and can’t come up with original prompts on your own, Meta says that Muse comes with “presets” — prefabricated image prompts — to “spark ideas.” However, a particularly eyebrow-raising feature allows users to manipulate another Instagram user’s images with AI, as long as that user’s profile is public. Users merely tag the person, and it allows them to take their picture and use it to create a new AI image.

This functionality has drawn sharp criticism from privacy advocates and users alike. One X user, after The Verge first pointed out how potentially invasive this is, stated: “Pulling real users into generated photos without explicit consent is a privacy landmine waiting to detonate.” Meta’s policy states that “people may be able to create content with your Instagram content using AI features at Meta” and that “you will not be notified about content created using AI features at Meta.” The company claims users “have control” over this feature, noting that there are settings you can use to disable this kind of co-option of your pictures if you want to.

Muse has other, less invasive applications. One is creating custom ads (AI has notably crept into advertising over the past year). Another is experimenting with interior decorating ideas — in a promotional video, a user leverages Muse to see what a secondhand couch might look like in their garage. This last function is designed to integrate with Facebook Marketplace, Meta’s popular Craigslist-like locus of used furniture and accessories. The model also features prompt-based image editing, which lets users create images to share across Meta’s apps and platforms.

“Ask it to mock up an image of you in front of a historical landmark, cleanly erase a photobomber from the background of a shot, or write a custom prompt to build a functional QR code,” the company offers. At the same time, Meta is launching a host of new AI effects for Instagram Stories, powered by Muse — notably, the same platform at the center of the photo-tagging concerns above. Those effects include customizable filters that can modify existing photos.

Meta says use of the new AI model is free for “everyday creation,” though users will need a subscription plan once they exceed a certain limit. The company also said Muse Video — presumably an AI video generator — is “already in development.”

Meta has released a number of AI apps and services over the past year, including an AI assistant called Creator, and Pocket, an app that can be used to vibe-code video games. The company has been accused of having a nebulous AI strategy, although it’s still on track to spend a whole lot on AI infrastructure this year as it continues to build out its services.

Meta’s privacy record is one reason for users’ unease over Muse. The company paid a then-record $5 billion fine to the FTC in 2019, after regulators found that the political consulting firm Cambridge Analytica had improperly harvested data from tens of millions of Facebook users — without their knowledge — to build voter-targeting profiles ahead of the 2016 U.S. election. Facebook had known about the data misuse for years before it became public. Separately, the company shut down Facebook’s facial-recognition system in 2021 — a tool that had automatically recognized people in photos and videos — amid lawsuits and regulatory pressure over its collection of biometric data. Basically, Muse’s photo-tagging feature, which is opt-out by default, fits a pattern users and regulators have flagged before: broad use of people’s data unless they actively turn it off.

The controversy over Muse Image is not isolated. It echoes past debates around deepfakes and consent in generative AI. Other companies, such as OpenAI with DALL-E and Stability AI with Stable Diffusion, have faced similar criticism when their models were trained on scraped images without explicit permission from creators or subjects. Meta’s approach, however, is unique because it directly targets social media users and their existing content on Instagram, making the privacy implications more immediate and personal. The opt-out nature of the feature places the burden on users to navigate complex settings, which many may not know exist or how to use. This has led to calls for stronger default privacy protections.

On the technical side, Muse Image is built on Meta’s foundational AI research, leveraging advances in large language models and diffusion models. The company has invested billions in AI infrastructure, including custom chips and data centers, to support such tools. Meta’s AI strategy has been described as scattered, with multiple products across different divisions, but the launch of Muse signals an attempt to unify consumer-facing AI under the Meta AI brand. The integration with Instagram Stories and WhatsApp suggests Meta sees AI as a key driver for user engagement on its platforms, which have seen slowing growth in recent years.

Despite the backlash, Meta is doubling down on AI-powered features. The company recently introduced an AI assistant called Creator that helps influencers generate content, and Pocket, an app for building simple games using natural language. Both have received mixed reviews. Muse Image, however, could be a turning point if the privacy concerns are not addressed. Regulators in Europe and the U.S. are already scrutinizing AI-generated content and deepfakes, and Meta’s history with data misuse makes it a prime target for enforcement.

Industry observers note that the photo-tagging feature could also enable malicious uses, such as creating non-consensual images or spreading misinformation. Meta has stated that it has safeguards in place, including content moderation and reporting mechanisms, but critics argue these are insufficient. The lack of notification to users whose photos are used is particularly troubling, as it removes the opportunity for informed consent. In response, Meta has pointed to existing privacy settings that allow users to limit who can tag them, but these settings were designed for traditional tagging, not AI manipulation.

The launch of Muse Image comes at a time when public trust in social media companies is low. Surveys show that a majority of Americans are concerned about how their data is used by tech platforms. Meta’s efforts to rebuild trust, such as its privacy-focused updates for WhatsApp and Messenger, may be undermined by features like Muse that appear to prioritize data usage over user control. The company has also faced lawsuits from states and individuals over its data practices, adding to the legal risks.

Looking ahead, Meta plans to expand Muse’s capabilities with Muse Video and possibly other modalities. The long-term vision is to create a comprehensive AI platform that spans text, image, video, and 3D content, all integrated with Meta’s social ecosystem. However, achieving that vision will require navigating a minefield of ethical and regulatory challenges. For now, users and advocates are watching closely to see how Meta responds to the backlash against Muse Image and whether it will adjust its default privacy settings to require explicit consent before photos can be used in AI generation.

Source:TechCrunch News