Facial recognition has made Wi-Fi security cameras much more useful, leading to more helpful notifications and even track who’s home. But that leads to some major privacy concerns, especially when these devices are used in public spaces.

See, facial recognition software can keep track of who goes where, and in places where privacy needs to be protected — like hospitals, banks and other sensitive institutions — that’s a problem. Israel-based startup D-ID (short for de-identification) is helping to protect privacy while keeping video viewable for humans.

The company alters video on a level that only artificial intelligence system notice, and the alterations cause AI tracking to think one face belongs to different people or isn’t even a face at all. But the security guard sitting in the video booth is still able to recognize a person they saw on the video screen.

The tech is software-based, altering the videos before they’re stored or analyzed by an AI, so you don’t have to worry about criminals foiling your home security cameras. But for institutions obliged to protect privacy, this allows them to stay secure with video recordings and maintain people’s freedom to anonymity.

The tech may also make it easier to share photos without worrying about your privacy. In fact, the company was created when the founders, who were serving in the Israeli military, couldn’t share photos for security reasons.

The obfuscation D-ID provides means people have to check photos by hand instead of handing them over to an AI — something only high-value targets really have to worry about. So you may soon be able to share that compromising college shot on Facebook without your boss getting a note about it.