Your public posts on Facebook are training Meta's AI without your knowledge

Have you ever wondered how tech giants train their powerful AI models these days? It turns out that your public posts on Facebook and Instagram have probably been sent to AI without your consent for a while now. You’re probably thinking — wait, how much of my stuff have they taken without my knowledge? And is there a way to stop them from using more or is it too late to pull the plug at this point?

According to ABC News, Meta, the tech giant behind Facebook and Instagram, has been quietly using our public posts and photos to train its artificial intelligence models for more than 15 years — since 2007.

This came to light during an Australian government investigation, with Melinda Claybaugh, Meta's global privacy director, initially denying the allegations but eventually confirming the practice under continued questioning.

What’s particularly troubling is the lack of transparency around this practice. Meta hasn’t been forthcoming about when it began collecting data or how long it has been going on. It recently confirmed that setting posts as private will prevent future data collection, but what about all those years of public posts? For many of us who joined these platforms in our teens or early twenties, the idea that our youthful indiscretions could be training an AI is, frankly, disturbing.

Even more troubling is the situation for users who were minors when they started posting. While Meta claims it doesn’t collect data from users under 18, it’s unclear how it handles accounts that were created by minors but are now adult accounts. This gray area raises serious questions about consent and data protection for young users.

According to a public statement from Meta, they rely on what they call “legitimate interests” as a legal basis for using public content to train AI. In simple terms, this means that they believe that using your public posts to improve their AI services can be justified as a reasonable business need.

It’s worth noting that this is a legally recognized concept in data protection law, specifically under the GDPR (General Data Protection Regulation), so Meta isn’t acting blindly here. However, it’s still a bit of a mystery — which may not be to everyone’s taste, especially when you consider how much of our lives we share online, both knowingly and unknowingly.

Now, if you're not keen on the idea of ​​your family vacation photos or smart story updates contributing to Meta's AI empire, you have an opinion on the matter.

You can opt out of AI training, but the process isn't always clear.

Opting out of AI training should be easy, but it’s not — especially if you live outside the EU. Meta doesn’t guarantee that it will let users opt out, but says it will “review objection requests in accordance with relevant data protection laws.”

If you live in the European Union, consider yourself lucky. Thanks to the General Data Protection Regulation, you have the right to object to how your data is used, and Meta is required to fulfill your request. All you have to do is fill out theObject to Meta's use of your information for AIFill out the form with your information using Facebook or Instagram.

The form optionally asks you to explain how Meta's data processing affects you. You can simply state that you wish to exercise your right under data protection law to object to the processing of your personal data. Enter your email address and you should soon receive an email and a notification on your Facebook account confirming that your request was successful.

However, for those in the US or outside the EU, where privacy laws are much less stringent, the path to opting out is more complicated. There’s no option to directly block Meta from using your public posts, comments, and even photos to train its AI. Instead, you have to take a few extra steps that feel more like a workaround than a solution.

To start, if you’re concerned about your data being collected to train AI, the first thing you can do is make sure your posts aren’t set to “public.” So by setting your audience to “friends” or something more private, you can at least prevent new posts from creeping into Meta’s dataset.

You can also set your Facebook account to private to minimize the risk as much as possible. Instagram also allows you to change your account to private.

But what about the data Meta has already collected? This is where things get a little more complicated. You can navigate to Meta’s Help Center to fill out a form titled “Data Subject Rights for Third-Party Information Used for Meta’s AI.” This form gives you three options: You can request access to your personal information that Meta uses to build or improve its AI, request its deletion, or raise concerns about your personal information appearing in response to Meta’s AI.

But Meta doesn’t make any guarantees, and will review each request according to local laws, meaning there’s no universal guarantee that your data will be completely erased. It’s not an opt-out solution per se, but it at least gives you more control over what happens to your data.

As I think about all this, I can’t help but wonder how things would have been different if we’d known a long time ago that our posts could help train AI. So, the next time you hit publish, maybe take a moment to think — are you comfortable with your words and images becoming the stuff of AI’s next big breakthrough?