For years we thought that the risk of Instagram was being watched too much. Now the problem changes shape: it no longer just concerns who sees our photos, who saves them, who comments on them or who shares them, but who can use them as raw material to create something that has never happened. A new scene, a new image, a new context. With inside, however, our face, our body, our public profile, our digital identity.
This is the most delicate point of the launch of Muse Image, the new image generation model presented by Meta, Mark Zuckerberg’s company that controls Facebook, Instagram and WhatsApp. Officially it is a step forward in creativity assisted by artificial intelligence: you write a prompt, ask the AI to create or modify an image, you get content ready to share in the apps of the Meta ecosystem. But inside the press release there is a sentence that has raised the alarm among users and privacy experts: it will be possible to mention Instagram accounts in the prompts and bring those profiles into the images generated by the AI.
In other words, if an Instagram profile is public, its photos can become a visual reference for a new artificial creation. It’s not the old repost, it’s not the tag in a Story, it’s not the screenshot of an already existing photograph. It is a further step: the image published years ago to describe a holiday, a job, an evening, a campaign, a family moment, can be used to build a different, plausible, realistic, but generated content.
The point, however, is not just privacy understood in the classic sense of the term. It’s no longer just about protecting a photo, preventing a download, avoiding an unauthorized repost or limiting the visibility of a profile. With generative artificial intelligence, something deeper comes into play: digital identity. That is, the set of images, gestures, poses, places, clothes, expressions and fragments of life that we have delivered to the platforms over the years and which today can become raw material to create new, artificial and potentially out of control versions of ourselves.
For this reason, Muse Image should not be read just as yet another creative tool launched by Meta. It should be read as a change of phase. Until yesterday, publishing a photo meant accepting that that photo could be seen, commented on, saved or shared. Today it can also mean allowing that image to contribute to generating something that we have never done, never said, never experienced, but which can resemble us enough to seem credible.
The feature that turns Instagram into an archive for AI
Meta introduces Muse Image as its first image generation model developed by Meta Superintelligence Labs and integrated into Meta AI. The system allows you to create images from scratch, modify existing photographs, combine visual references, use creative presets and obtain content to download or share directly in chats, Stories or the feed.
The most discussed function, however, is the one linked to Instagram accounts. Meta explains that users can “@mention” Instagram profiles in the Meta AI app to bring them into their images. According to the company, tagging a username allows the AI to use public photos to build a ready-to-post visual.
This is where creative promise becomes a matter of control. Until yesterday, a public profile meant accepting that anyone could see those images. Today the boundary is moving: public does not only mean visible, but potentially reusable by a generative system. And the difference is enormous, because it’s one thing to show a photo, it’s another to allow that photo to contribute to creating new content.
Because Muse can become a problem
The problem is not only technological, but cultural. Instagram was built on the idea that images told something about us: private life made aesthetic, work made recognizable, social presence transformed into a visual archive. With generative artificial intelligence, that same archive can change function. It no longer serves only to represent us, but also to recreate us.
For creators, influencers, journalists, artists, professionals and more generally anyone with a public profile, the point is clear. The face, style, clothes, poses, places frequented and the way a person presents themselves online can become elements to combine in a new image. Even when the result does not explicitly violate the platform’s rules, a gigantic question remains: who decides in what context our identity can appear?
Meta says Muse Image is built with security controls and barriers, and that content that violates community standards can be removed or reported. But the concern is not just about the most extreme abuses. It’s about the normalization of a principle: if your profile is public, the system can treat it as available creative material, unless you decide otherwise.
The real problem is not the photos, but the digital identity
Our digital identity is no longer just the profile we build online. It has become a repository of visual data for platforms to work with. Every public photograph says something: how we dress, how we smile, where we go, who we are with, what style we have, what objects surround us, what environments we frequent. Taken individually, these details seem innocuous. Put together, however, they make up a precise representation of a person.
Generative AI changes how that representation can be used. It doesn’t just recover an existing image, but can produce a new one. It doesn’t simply show what we have published, but can imagine something else starting from what we have published. This is where the issue becomes delicate: the boundary between online presence and identity simulation is becoming increasingly thin.
For a public figure, a journalist, a creator, an artist or a professional, the risk is evident. But the topic concerns everyone. Because over the years we have all, to varying degrees, built a digital double made up of photographs, tags, places, expressions, habits and relationships. And that double, today, can be reused by tools that don’t just tell us: they can recreate us.
This is why talking about “cloning” of the personal image, even with all the appropriate quotation marks, conveys the idea better than many technical formulas. This is not necessarily a perfect copy, nor a deepfake in the strict sense. However, it is about the possibility of using a public profile as a reference to produce new images, in which a person can be evoked, reconstructed, inserted into artificially generated scenarios. And in an ecosystem where images circulate faster than denials, even a credible resemblance can be enough to create confusion.
The opt-out issue
And this is precisely the most controversial point. Protection does not arise as prior consent, but as a subsequent renunciation. If the user does not want their public content to be used for this type of AI function, they must manually intervene in the settings.
The logic is that of the opt-out: you are not asked first if you want to participate, you have to discover the existence of the function and deactivate it. It is a mechanism that has become familiar in many digital platforms in recent years, but which in the case of visual identity produces a particularly sensitive effect. Because we’re not just talking about abstract data, preferences, histories or interactions. We are talking about faces, bodies, people, children possibly present in public photos uploaded by adults, private environments that have ended up in an open profile over the years.
According to international media reports, private accounts and those belonging to minor users would be automatically excluded. But this does not close the problem. On the one hand, many adult users maintain public profiles for work, reputation or simple habit. On the other hand, a minor can still appear in public photos of an adult, for example in family, school, sports or holiday images.
The issue, therefore, does not only concern those who use Instagram for work or those who have thousands of followers. It concerns anyone who has left a public visual trace online over time. Because digital identity is not only made up of the contents we publish today, but also of those we forgot we published yesterday.
Those who have a private account should not have to do anything at the moment: the photos are not public and should not be available for this type of reuse. Those who have a public profile and don’t want to make it private should check the app settings.
The path indicated is this: you need to open Instagram, go to your profile, touch the menu with the three lines at the top right and look for the “Sharing and reuse” section. Inside, an entry should appear linked to the possibility of allowing other people to reuse their content on Instagram and with Meta’s AI functions. At that point the toggles relating to posts and Reels must be deactivated.
The wording may change depending on the app version and country, and not all users may see the setting immediately, because the rollout of the feature is progressive and starts in the United States and selected countries. Precisely for this reason, however, the advice is to check periodically: the fact that the option does not appear today does not mean that it cannot appear in the coming weeks.
There is another important detail. Turning off the setting serves to prevent future use, but does not necessarily delete any AI images already created before opting out. Protection is therefore not retroactive. This also helps to make the topic less banal than it may seem.
It’s no longer enough to wonder who sees our photos
Meta’s new feature comes at a time when all the big platforms are trying to make generative AI an everyday habit. No longer an external service to be opened when needed, but a presence incorporated into the apps we already use for messages, photos, videos, purchases, advertising, work and entertainment.
The point, though, is that generative AI doesn’t just organize what exists. He reworks it. He puts it back together. He imitates him. It transforms it into something new. For this reason the decisive question is no longer just “who can see my photos?”, but “what can be done with my photos?”.
Instagram has always lived on an ambiguous tension between exposure and control. Publishing meant making yourself visible, but within a recognizable perimeter: your profile, your feed, your audience. Now that perimeter becomes more porous. A photograph does not necessarily stay where we put it. It can become an element within a prompt, a tile within a composition, a face within a generated scene.
This is the real news. Not the fact that there is a new image generator, because now every big tech company has one. The news is that the raw material of that generator can be our life published online, and that to avoid it you need to know where to look.
The point is not to be afraid of artificial intelligence. The point is to understand that the protection of one’s image online can no longer stop at the simplest question: “who can see my photos?”. The new question is another: “what can be generated from my photos?”.
This is where Muse Image becomes bigger news than the single feature. Because it tells of the transition from a web in which our identity was observed to a web in which it can be reconstructed. And, perhaps, this is precisely the new frontier of privacy: not just deciding what to show, but maintaining control over what can be created using us as a starting point.
For those who have a public Instagram profile, the advice is simple: check the settings immediately. Not out of panic, but out of awareness. Because privacy, in the age of artificial intelligence, is no longer just about what we decide to hide. It also concerns what we have already shown, perhaps years ago, without imagining that one day it could become an image never taken.
