All About Cookies is an independent, advertising-supported website. Some of the offers that appear on this site are from third-party advertisers from which All About Cookies receives compensation. This compensation may impact how and where products appear on this site (including, for example, the order in which they appear).
All About Cookies does not include all financial or credit offers that might be available to consumers nor do we include all companies or all available products. Information is accurate as of the publishing date and has not been provided or endorsed by the advertiser.
The All About Cookies editorial team strives to provide accurate, in-depth information and reviews to help you, our reader, make online privacy decisions with confidence. Here's what you can expect from us:
One byte was enough to switch off the safeguard meant to stop unapproved code from running on Meta smart glasses. No stolen password, sophisticated hacking toolkit, or weeks of brute-force guessing.
These are the same glasses millions of people wear every day. They have a camera that sees what you see and microphones that hear what you hear. And they sit inches from your eyes for hours at a time.
Earlier this month, security researcher Paula (known online as pamoutaf) demonstrated that she could install and run her own code on Meta’s smart glasses. Meta later introduced a fix, but she found a way to switch it off.[1]
Here are more details about her findings and what you can do to protect yourself from unwanted and non-consensual monitoring.
The glasses trust the phone they’re paired with, and that trust is what Paula exploited. Here’s how she went from a borrowed pair of glasses to typing commands into them from her phone.
The first step in Paula’s process was to get the firmware. Paula decompiled the Meta AI Android app, which is the companion app for Meta glasses, using a tool called JADX, which turns a compiled app back into readable code.
She then used Frida, a security research tool that allows you to hook into and change a running app’s behavior. Using it, she pulled out the over-the-air (OTA) update download URL that the app requests from Meta’s server.
The incremental update file was about 156 MB, which only patched an existing system. In the app’s code, she found a GraphQL request field called device_firmware_full_update_only, which is normally set to false. She forced it to true and got the full firmware image.
Meta’s developer platform, the Device Access Toolkit (DAT), works by having the phone download a DAT APK and install it on the glasses. Up to firmware version 126, the glasses did not verify whether this APK was signed by Meta.
Due to this flaw, Paula was able to intercept the installer in the Meta AI app and replace Meta’s APK with a self-signed APK of her own. The phone delivered it, and the glasses installed it.
The DAT app only starts when a developer app opens a DAT session. Paula used Meta’s own sample app, Camera Access, to trigger a session. For debugging, she used Android’s tone generator to make the glasses beep in coded patterns.
Three beeps meant the listener had started looking for a Wi-Fi Direct address. Five beeps meant the address had been found. Four beeps meant that her listener was ready, and six beeps indicated that something was wrong.
When syncing photos and videos, the Meta AI app creates a Wi-Fi Direct network (in the 192.168.49.x range) that the glasses join. Paula’s app opened a listener on TCP port 31337 on that network and used Netcat (nc) on her phone to run commands.
However, about five days later, her access disappeared, and firmware version 127 began enforcing signature verification of the DAT APK. The on/off switch for this check lived in a settings file called MobileConfig, which the phone fetches from Meta and passes to the glasses.
The glasses checked the file only for structural validity, not its integrity or authenticity. Paula edited the binary file holding the values, changing one byte. The glasses accepted this change, skipped the signature check, and installed the self-signed APK again.
The good news is that this hack isn’t something anyone can pull off remotely. This method works only from the phone paired with the glasses. Whoever controls the Meta AI app controls what gets sent to the glasses. Plus, the researcher’s code only ran with limited, ordinary app-level permissions, not full control of the device.
Paula hasn’t demonstrated that it could access the Meta glasses’ camera, microphone, or stored recordings. Malwarebytes says that this might be a possibility. Paula reported the flaw through Meta’s bug bounty program. Meta approved publication of her research, which she takes to mean that it has been fixed.
For most owners, the realistic risk comes from someone who gets into their phone. This could happen through malware with remote access capabilities or through a person who has access to their unlocked device.
This is why it’s important to keep the firmware and Meta AI app updated and avoid unofficial apps at all times. Also, be aware of phishing links that look convincing. When you click on them, they could install malware that gives an attacker control of your device and, consequently, the Meta AI app.
These glasses carry a camera and microphone and sit on your face all day. The researcher demonstrated that it was possible to bypass a security check simply by changing one byte in the file. That’s worrying, given what Meta plans to do next with the glasses.
Meta filed a patent for an AI assistant that watches through the glasses’ camera, identifies interesting moments, and automatically captures and compiles them into a highlight reel. The patent also proposes using facial recognition to identify people and distinguish close friends or relatives from mere acquaintances.
Some pubs in the UK have already banned Meta AI glasses, citing privacy concerns. The main concern is that the glasses look like ordinary eyewear and can capture the faces, voices, and conversations of people without their consent.
Paula’s findings have underlined the privacy risks surrounding Meta smart glasses once again, adding to concerns about their growing popularity as creepy glasses. The research shows that someone who controls the paired phone could, at least before Meta's fix, install their own software on the glasses.
However, the good news is that this isn’t a remote hack that anyone could execute from a distance. They would need access to the phone linked with the glasses and control over the Meta AI app. To stay safe, ensure that you update the firmware and Meta AI app regularly. Do not click on phishing links that may install malware on your device.
Top-rated data removal service that scrubs your info from 420+ data broker sites automatically
Independently verified by Deloitte, meaning removals are actually sent, confirmed, and not just claimed
The Unlimited plan extends coverage to 2,000+ additional sites with human-assisted custom removals
/images/2026/06/04/best_lifelock_alternatives.jpg)
/images/2023/10/12/best_identity_theft_protection_with_dark_web_monitoring.jpg)
/images/2026/11/13/meta-smart-glasses-hack.png)
/images/2026/11/13/fbi_seized_900_secretly_recorded_webcam_videos_from_a_fake_zoom_call_scam.png)
/images/2026/11/12/mobile_police_wont_say_how_they_use_clearview_ais_100_billion_face_photos.png)
/images/2026/11/09/sony-ad-scent-patent.png)
/images/2026/11/09/plex-server-security-flaws.jpeg)
/images/2026/11/06/mamdani_wants_to_stop_madison_square_garden_from_scanning_your_face.png)
/authors/krishi_chowdhary_headshot.jpg){kind=small}
/authors/kalleigh-lane-new.jpg){kind=small}
/images/2025/09/11/incogni-logo.png){kind=logo}
/authors/krishi_chowdhary_headshot.jpg)