This $150 mask is made up of 2D images, a silicone nose, a specially processed area, and a 3D printed frame. Other parts of the mask are just 2D images and a makeup artist has done a bit of work too. As can be seen in the video below, it appears to trick Face ID on the first attempt. Each capture hones the AI's ability to distinguish the owner from an imposter.
The mother, Sana Sherwani, and the son, Ammar, realized the Face ID could be fooled shortly after purchasing the device, Attaullah Malik, the father of Ammar, said in a LinkedIn post. "These are actual masks used by the engineering team to train the neural network to protect against them in Face ID".
As far back as 2009, Bkav researchers highlighted what they said were problems with using facial recognition as a way to authenticate users.
Bkav also published a YouTube video that shows one of their staff members pulling a piece of cloth from a 3D-printed mask. The firm used different materials to create certain portions of the mask. It only took Vietnam-based cybersecurity experts from Bkav a few days to figure out the vulnerabilities of the authentication mechanism and create an extremely creepy mold of a hacker's face. "This seems like an unlikely sequence of events". The FaceID failed to tell the difference and unlocked the phone. Is your data so valuable that someone would go to this effort? "Exploitation is hard for normal users, but simple for professional ones". The iPhone X will still recognise you even when half of your face is covered, the company says. Apple's facial recognition begins with the opening assumption that the user gazing at the screen is likely to be the correct user.
"It's not easy for normal people to do what we do here, but it's a concern for people in the security sector and important people like politicians or heads of corporations", he said.
Although many have viewers also called out the authenticity of the experiment.