Meta’s New AI Image Model: A Double-Edged Sword
Meta has recently unveiled its new Muse Image model, the first release from the ‘Meta Superintelligence Labs.’ This cutting-edge technology promises to create hyperrealistic images, posters, and even videos—once Muse Video launches, that is. While this advancement is certainly noteworthy, it raises some pressing concerns regarding user privacy.
As Wired reports, there’s something peculiar about how Meta’s new AI image generation feature integrates with public Instagram profiles. By default, anyone using Meta AI can create an image using the likeness of a public Instagram account, which can then be shared via chat, Stories, or the Instagram feed. This doesn’t have to be a celebrity or a famous person, either; in fact, some celebrity profiles are restricted from this feature. So long as it’s public, your personal Instagram profile is fair game here.
This is a significant privacy issue. Every public profile is automatically opted into this new feature without their consent. If you’re using Meta AI features on Instagram, you’re in on this as well. For Meta, the distinction is clear: if you’re posting anything publicly, they can use it on their AI platform. Of course, creating an image out of two different faces is nothing new. You can just upload two profile images and get the same effect. But Meta is making this trivially easy on a very large scale.
I tested this out in a couple of ways. First, I attempted to make an image of myself in a fight scene with Robert Downey Jr. The AI refused, possibly due to image rights. Then, I asked Meta AI to create an image with myself and my colleague Pranay Parab. That wasn’t possible because he has a private profile. Lastly, I asked it to create an image of myself having coffee with my friend Dhaval, who owns a cafe and has a public profile. That image was created in just under a minute and was based on recent images uploaded by Dhaval and me on our respective profiles (both a couple of years old at this point). It’s clear that this technology is easy to use, but can you turn it off?
If you have a private Instagram profile, this isn’t something you need to worry about—beyond mutual friends taking your photos for their own AI usage. If you have a public profile, you need to disable a setting that lets Meta use your profile data for Meta AI and Reels generation. Here’s how:
How to Stop People from Using Your Instagram Profile in Their AI Images
To disable this feature, open the Instagram app on your iPhone, go to your Profile, and tap the three-lined Menu button in the top-right corner. Then, go to Sharing and Reuse and navigate to the section called ‘Allow people to reuse your content on Instagram and with AI features at Meta.’ Here, disable the ‘Posts and Reels’ feature.
The good news is that this does work. I tried it on my Instagram coffee page that I rarely use. When I was logged into my personal profile on Meta, I asked Meta AI to create an image of me brewing coffee using my coffee page handle, and it refused, saying it doesn’t have the right to do that. This, of course, should be the default.
It’s worth noting that this integration with public Instagram profiles is a stark contrast to Google’s similar feature, which is limited to the user themselves and requires an approval process. The lack of guardrails in Meta’s approach raises significant concerns regarding user consent and data privacy.
In conclusion, while Meta’s new AI image model is certainly an impressive achievement, it’s essential to address the pressing concerns surrounding user privacy. By disabling the necessary settings, users can prevent their public profiles from being used in AI-generated images without their consent.
For those who are concerned about their public profiles being used in AI-generated images, it’s essential to take action and disable the necessary settings. By doing so, you can ensure that your personal data and likeness are not used without your consent.
