In a blog post on Tuesday, Meta announced the release of an AI image generator called Muse Image, along with an as-yet unreleased video generator called Muse Video. This move places Meta in a competitive landscape already populated by similar tools from Elon Musk’s X (formerly Twitter) and OpenAI. X rolled out an image generator with social media integration about two years ago, while OpenAI released its video generator via the Sora social media app in September of last year. However, OpenAI later announced it was killing Sora this past March before pivoting the company’s entire focus. Meanwhile, X’s integration of Grok images has been fraught with controversy. During the holiday season, it became a meme on X to ask Grok to generate sexualized images of people, including minors, raising serious ethical and legal questions.
Given these precedents, one might assume Meta would tread carefully with Muse Image. Instead, the company appears to have doubled down on the same approach: pulling user data from Instagram to fuel its AI generation. If you have a public account, anyone using Meta AI can generate images of you without your explicit consent. This default setting has alarmed privacy advocates and casual users alike.
How Muse Image Works
To see Muse Image in action, you can go to the Meta AI web app at meta.ai, log in using your Instagram account, and ask the chatbot to generate a picture of yourself. In tests conducted by multiple reporters, the tool attempted to scrape photos from the user’s own feed. For users with private accounts, the system might still attempt to access photos if you are logged into that account. However, the tool’s chain-of-reasoning display may indicate insufficient data, resulting in a default image of a random person instead.
More worryingly, users can also prompt the chatbot to generate an image using details pulled from someone else’s public Instagram account. In one test, a reporter was able to generate an image of Mark Zuckerberg, and even more concerning, they generated an image of a real-world friend they had never followed or interacted with on Instagram—essentially a digital stranger—without asking that person’s permission. This raises serious questions about identity theft, harassment, and the potential for deepfake misuse.
Default Settings and Privacy Implications
As noted by technology analysts, allowing others to access your likeness in this way is currently the default setting in Instagram if you have a public account. Instagram has updated its relevant help page to disclose this. The section now states: “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, videos or reels in features like remix, sequence, templates and stickers. In addition, people may be able to create content with your Instagram content using AI features at Meta. Depending on the settings of the other user, this means your reused content may be discoverable in search engine results.”
This means that any public photo you’ve posted could be used to generate entirely new AI content without your knowledge or consent. The potential for abuse is significant: bad actors could create compromising images, spread misinformation, or even impersonate you. While Meta claims to have built-in protections to prevent policy-violating content—such as violent, sexual, or defamatory imagery—previous incidents on other platforms have shown that such safeguards are often insufficient.
How to Opt Out of Muse Image
If you are uncomfortable with your likeness being used in AI-generated content, there are steps you can take to opt out. The most drastic and effective method is to set your Instagram account to private. This prevents anyone who does not follow you from accessing your photos and videos, thereby blocking the AI tool from using your content. However, this also limits the reach and engagement of your account, which may not be ideal for influencers or businesses.
Alternatively, you can adjust your settings directly in the Instagram app. Navigate to your profile, tap the hamburger button in the top-right corner, and find the tab for “Sharing and reuse.” There you will see an option labeled “Allow people to use your content on Instagram and with AI features on Meta.” Toggle both “Posts” and “Reels” to the off position. This setting appears to be available for U.S. users on the smartphone app, and according to Meta’s blog post, similar functionality is being rolled out to Instagram Stories for U.S. users and in some territories to WhatsApp. Facebook will reportedly get the option soon.
It is important to note that even after opting out, any content that has already been used to train AI models or generate images may still persist. Meta has not provided clear details on whether opting out will retroactively remove your data from existing models.
Historical Context of AI on Social Media
The launch of Muse Image comes at a time when public trust in social media companies regarding data privacy is at an all-time low. The Cambridge Analytica scandal in 2018 exposed how Facebook (Meta) allowed third-party apps to harvest user data without consent. Since then, regulators worldwide have tightened rules, most notably with the European Union’s General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA). However, the rapid rise of generative AI has outpaced legislative efforts.
OpenAI’s Sora, which offered AI-generated video from text prompts, faced similar privacy criticisms before being shut down. X’s Grok image generator led to widespread misuse, including generating inappropriate images of real people. In both cases, the companies responded by updating their policies and adding moderation tools, but the damage to user trust was already done. Meta seems to be following the same playbook, launching a powerful tool first and addressing concerns later.
Another key difference is that Meta already possesses a vast repository of personal data from its billions of users across Instagram, Facebook, and WhatsApp. Integrating AI image generation means that data can be repurposed in ways users never anticipated. Privacy advocates have called for Meta to make opt-out the default, rather than opt-in, and to provide clearer disclosure about how user data is used in AI training.
Potential Risks and Mitigations
The ability to generate realistic images of real people without permission opens the door to several risks. Identity theft is a primary concern: a generated image could be used to create fake social media profiles or impersonate someone in phishing scams. In the wrong hands, Muse Image could also be used to create non-consensual intimate imagery, a form of abuse that has already plagued other platforms. Additionally, deepfake images could be used to spread false information, damaging reputations and even influencing elections.
Meta has acknowledged these risks. In a statement provided shortly after the tool’s publication, a Meta spokesperson said: “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 argue that automated moderation systems are not foolproof, and that the burden should not be on users to report violations after they occur.
Furthermore, the same statement does not address the broader privacy infringement of using someone’s likeness without consent even for benign purposes. Civil liberties groups have already begun exploring legal avenues to challenge AI platforms that scrape public data without explicit opt-in. In Europe, GDPR may already require a lawful basis for processing such data, which could put Meta in legal jeopardy.
Comparative Analysis with Other Platforms
X’s Grok image generator, launched two years ago, allows users to create images based on prompts that can include references to real individuals. The platform faced immediate backlash when users generated sexually explicit images of public figures and minors. In response, X updated its terms to prohibit such use and added automated filters, but enforcement has been inconsistent. OpenAI’s new version of DALL-E also blocks requests involving real public figures by default, but Meta has not implemented such a safeguard for Muse Image.
Unlike these competitors, Muse Image is deeply integrated into Instagram’s social ecosystem. The ability to generate images that incorporate your friends’ faces or backgrounds is a feature designed to encourage viral content creation. Yet the same integration makes it easier for users to create harmful content without understanding the implications. Meta has not released detailed data on how many users have already been affected or how many images have been generated using real people’s likenesses.
What Users Can Do Now
Given that Muse Image is now active for U.S. users on Instagram, anyone concerned about their digital privacy should take immediate action. The fastest way to limit exposure is to set your account to private. If that is not feasible due to professional or personal reasons, at least disable the “Sharing and reuse” toggle described above. Additionally, consider reviewing your past posts and removing any that could be particularly sensitive—such as photos of your home, family members, or identification cards.
For parents, it is especially important to lock down children’s accounts, as minors are more vulnerable to exploitation. Meta’s terms require users to be at least 13 years old, but enforcement of age verification is minimal. If an account is public, any photo of a minor could be used in AI generation, potentially creating risks that parents may not have anticipated.
Beyond individual action, advocacy groups are pressuring Meta to implement stronger defaults. A petition started by the Electronic Frontier Foundation (EFF) calls on Meta to make the opt-out the default for all users, not just those who know about the setting. They argue that informed consent is impossible when the feature is buried in menus and not prominently disclosed at the moment of posting. Meanwhile, lawmakers in the U.S. Senate have introduced a bill that would require social media platforms to obtain explicit consent before using user data in AI models, though its passage remains uncertain.
As Muse Video prepares to launch, the stakes will only get higher. Video generation is computationally more expensive but also more dangerous, as realistic video clips can be used for blackmail, propaganda, or fraud. Meta has stated that similar safeguards will be in place, but history suggests that the arms race between AI capabilities and safety measures will continue. Users must remain vigilant and proactive in protecting their digital identities.
Source: Gizmodo News