Meta’s long-awaited entry into the AI image generation market, Muse Image, arrived this week with a feature that may surprise many Instagram users: it can scrape your public photos to generate new images of you—without your explicit consent. The tool, unveiled in a Tuesday blog post, is integrated into Meta’s AI chatbot and allows anyone logged into Instagram to prompt the system to create a picture of any public account holder. While Meta insists it has built-in safeguards against violent, sexual, or defamatory content, the default setting for public accounts enables this reuse, prompting privacy experts to call for immediate opt-out actions.
Muse Image joins a growing list of social media–integrated AI generators, but it arrives well after competitors like Elon Musk’s Grok on X or OpenAI’s short‑lived Sora video app. Grok faced heavy criticism last holiday season when users generated inappropriate images of minors, and OpenAI shut down Sora in March after similar problems. Meta is attempting to avoid those pitfalls by employing what it calls “chain‑of‑reasoning” filters, but the core functionality remains controversial: the AI can access and remix real people’s likenesses from Instagram without their direct approval.
In tests conducted by this publication, a journalist with a private Instagram account found that Muse Image could still search his own photos when logged in, but failed to produce an accurate image due to insufficient data—instead generating a random person and labeling it as a placeholder. More concerningly, the tool successfully created images of Mark Zuckerberg and of a random friend who uses a public profile, even when the prompter had never interacted with that person on Instagram. The only barrier is that the target must have a public account; private accounts remain invisible to the generator unless the requester is also logged into that account.
The implications for public figures, influencers, and ordinary users are significant. According to Instagram’s updated help page, publicly posted photos, videos, and reels can now be reused by others using “AI features at Meta,” and depending on the other user’s settings, the generated content may even appear in search engine results. The language was added quietly alongside the launch of Muse Image, and many users have not received direct notifications.
Meta claims that its technology includes protections to prevent policy‑violating outputs. A spokesperson told Gizmodo: “Muse Image has built‑in protections to help prevent the generation of policy‑violating content, including violent, sexual, or defamatory imagery of real people. Content that violates our policies—whether reported by users or detected by our systems—is subject to enforcement under our Community Standards.” However, critics point to the Grok and Sora examples as evidence that such protections are often insufficient once a tool is in public hands.
The rollout of Muse Image comes at a time when generative AI is rapidly transforming social media. TikTok, Snapchat, and other platforms have introduced AI stickers and filters, but few allow users to generate entire photorealistic images of strangers based solely on their public feed. Meta’s integration goes further by linking the generation process to its core social graph, making it trivial to create images of anyone from celebrities to distant acquaintances.
For those who wish to opt out, the process is straightforward but not immediately obvious. The quickest method is to set an Instagram account to private, which automatically prevents Muse Image from accessing the profile’s photos. For users who prefer to keep their account public, a separate toggle exists: in the mobile app, go to your profile, tap the hamburger menu (three horizontal lines) in the top‑right corner, select “Sharing and reuse,” then turn off “Allow people to use your content on Instagram and with AI features on Meta” for both “Posts” and “Reels.” The same option is available in a desktop browser through the privacy settings.
While Muse Image is currently limited to image generation, Meta has announced that a video generator called Muse Video is coming soon. It will likely operate under the same data‑sharing framework, amplifying the privacy stakes. For now, the technology is only available in the United States for Instagram Stories, with Facebook integration promised in the coming weeks.
The broader context of this launch is Meta’s aggressive push into generative AI despite past controversies. The company’s own history with face‑recognition technology is checkered; it shut down a facial‑recognition system in 2021 after regulatory pressure and lawsuits. Some privacy advocates argue that Muse Image represents a workaround to those restrictions—using photos as training material for AI rather than for direct identification.
Industry observers note that the default opt‑in for public accounts mirrors a trend in social media: platforms increasingly assume user consent for new features unless explicitly revoked. “It’s a classic case of privacy by default,” said one digital rights analyst. “The burden is on the user to find the toggle, which most people will never do.”
As of now, no major privacy lawsuits have been filed against Meta over Muse Image, but the Electronic Frontier Foundation and other groups have issued warnings. Users are advised to review their Instagram privacy settings immediately, especially if they have a public account with identifiable photos. The ability for anyone—including strangers—to generate an AI image of you without permission poses risks ranging from deepfake harassment to reputational damage.
In the coming weeks, Meta plans to extend Muse Image to WhatsApp in select territories, further broadening the reach of the feature. Whether the company will respond to backlash by changing the default settings remains uncertain. For the moment, the power to protect your digital likeness lies in your own hands—provided you know where to look.
Source:Gizmodo News