Two Harvard students developed I-Xray, a program for Meta’s smart glasses that uses facial recognition to identify individuals and access personal information, raising significant privacy concerns.
Two Harvard engineering students have developed a program, I-Xray, for Meta's Ray-Ban smart glasses that uses facial recognition to identify individuals in public and access personal information such as names, phone numbers, and even home addresses. The program, created by AnhPhu Nguyen and Caine Ardayfio, taps into PimEyes, an AI-powered facial recognition tool, to match captured images with publicly available data. Once identified, the program pulls further details from various online databases, providing users with comprehensive information about the individual.
In a demonstration video, the students showed how I-Xray works by identifying classmates and strangers on the street in real-time. However, they clarified that their intent is to highlight the significant privacy concerns surrounding the use of Meta's smart glasses rather than release the tool for public use. "Some dude could just find some girl’s home address on the train and just follow them home,” Nguyen told 404 Media, emphasizing the potential dangers.
Meta, in response, stated that the Ray-Ban smart glasses do not include facial recognition technology and that similar privacy risks exist with any photo or video recording device. The glasses do, however, have an LED light to indicate when they are recording, and Meta has implemented tamper detection technology to prevent users from disabling this light. The company insists that it takes privacy seriously, although concerns remain over the capabilities of such wearable technology when combined with external facial recognition tools.
Nguyen and Ardayfio are urging individuals to take proactive steps to protect their privacy by removing their data from public facial recognition databases like PimEyes and opting out of people search engines. While this may offer some protection, experts warn that the potential for misuse of this type of technology is vast, particularly in an age where personal data can be accessed and exploited with ease.