On Tuesday, Meta announced the release of its long‑awaited AI image generator, Muse Image, along with a forthcoming video generator, Muse Video. While Elon Musk’s X introduced a similar social‑media‑integrated image generator about two years ago, and OpenAI launched its Sora video generator in September 2025, Meta’s entry feels both overdue and problematic. The company’s tool pulls data directly from Instagram, and for users with public accounts, it can generate images of anyone—without their explicit consent. This move has sparked immediate concerns about privacy, identity theft, and the potential for misuse.

How Muse Image Works

Muse Image is accessible through the Meta AI web app at meta.ai. After logging in with an Instagram account, users can prompt the chatbot to generate an image of themselves. If the account is private, Meta’s AI attempted to dig through personal photos anyway, though in one test it claimed insufficient data and created a generic image. However, the real shock comes when you ask it to generate an image of someone else. The tool can pull publicly available Instagram photos to create new pictures of acquaintances, friends, or even digital strangers with public profiles.

In tests conducted by journalists, prompts like “generate an image of Mark Zuckerberg” worked seamlessly. Even more concerning, generating images of random Instagram friends—people the user follows or even those with no direct connection—was possible without their permission. Meta’s default settings allow this behavior for any public Instagram account. The company updated its help page to disclose that, “If you have a public account, other Instagram users may be able to create new reels, posts or stories that reuse part or all of your published photos … In addition, people may be able to create content with your Instagram content using AI features at Meta.”

Historical Context: AI Image Generators on Social Media

Meta’s move comes after several high‑profile failures in the social AI space. X’s integration of Grok images allowed users to generate sexualized images of minors during the 2025 holiday season, sparking outrage and regulatory scrutiny. OpenAI, meanwhile, announced the shutdown of Sora in March 2026 after struggling with content moderation and ethical boundaries. Meta appears to have learned little from these precedents. The company claims it has built‐in protections to prevent violent, sexual, or defamatory imagery of real people, yet the very architecture of Muse Image—which scans public Instagram accounts—invites abuse.

Privacy advocates have long warned that aggregating personal photos into a generative AI training set without explicit opt‑in violates user trust. The Federal Trade Commission and European data protection authorities have investigated similar practices. Meta’s approach mirrors its past controversies, such as the Cambridge Analytica scandal, where user data was harvested without proper consent. The difference now is that the data is used not just for targeting ads, but for creating entirely new digital likenesses that can be manipulated, shared, and potentially used for misinformation.

How to Opt Out

Fortunately, users can disable this feature. The quickest method is to set your Instagram account to private. For those who prefer to stay public, a manual opt‑out exists in the smartphone app: go to your profile, tap the hamburger menu (three lines) in the top‑right corner, select “Sharing and reuse,” and toggle off “Allow people to use your content on Instagram and with AI features on Meta.” This removes both posts and reels from the pool of data that Muse Image can access. However, the process is buried in settings, and many users may never find it. Meta does not proactively notify users when enabling such features, a criticism that has been leveled against the company for years.

Muse Image is currently available to U.S. users on Instagram Stories and is expanding to WhatsApp in some regions. Facebook integration is expected soon. Meanwhile, Muse Video remains in development, though early reports suggest it will operate on the same data‑scraping principles. Without a clear, default opt‑out mechanism, millions of people may already have had their photos used to train or generate AI images without their knowledge.

Broader Implications for AI and Privacy

The launch of Muse Image highlights a growing tension between innovation and individual rights. Generative AI models thrive on vast datasets, and social media platforms are treasure troves of rich, labeled visual data. Meta’s decision to enable this by default reflects a corporate strategy that prioritizes data collection over user autonomy. It also sets a dangerous precedent: if Meta can scrape Instagram photos to generate images, what prevents other companies from doing the same? The legal landscape remains fragmented, with the European Union’s AI Act imposing stricter rules but the U.S. lacking comprehensive federal privacy legislation.

Another concern is the potential for deepfake abuse. While Meta asserts that its safeguards block explicit content, third‑party tests have shown that developers can often circumvent such protections. The history of AI image generators is filled with examples of “jailbreak” prompts that bypass content filters. Moreover, the generation of harmless images today could easily evolve into more malicious uses tomorrow, especially as the model improves. The fact that Muse Image can create realistic images of strangers without their consent opens the door to harassment, impersonation, and identity fraud.

Cybersecurity experts recommend that users review their Instagram settings immediately. Even if you trust your friends, the data you share may be used by third parties you never account for. In an era where AI can copy your likeness with a few keystrokes, the concept of public visibility itself must be reconsidered. The shift from passive sharing to active generation of user content represents a paradigm change that most people are not prepared for.

Meta’s spokesperson defended the feature by stating, “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.” Yet enforcement after the fact is poor consolation for someone whose face appears in an unwanted AI‑generated picture. The burden of monitoring and reporting falls on victims, not on the company that enabled the misuse.

As Muse Image rolls out more broadly, users should demand greater transparency and control. The “sharing and reuse” setting is a step in the right direction, but it should never have been the default. Privacy professionals advise regularly auditing your social media settings, especially after platform updates. For now, the simplest safeguard is to make your Instagram account private—or to pause sharing until Meta proves it can handle this responsibility ethically.

Source: Gizmodo News