Two students from the University of Harvard have recently done demonstration of worrying use of Ray-Ban smart glasses Metacapable of revealing the identity, addresses and telephone numbers of individuals in real time using facial recognition and public databases. This project, called I-XRAYshows how widely available technologies can be combined to circumvent privacy.
In a video posted online, AnhPhu Nguyen and Caine Ardayfio, the students behind the project, explain how their technology works. Ray-Ban glasses Metacapable of broadcasting live videos on Instagram, are associated with a computer program using artificial intelligence to identify the faces of the people filmed. These images are then analyzed by public databases, revealing personal information such as names, addresses and telephone numbers. The information is then sent back to a mobile app for viewing.
A chilling demonstration
The video shows the two students using their technology to identify multiple classmates in real time, harvesting their personal information. They are also seen approaching strangers on public transport, pretending to know them using information extracted via I-XRAY. The device, based on already available tools, highlights the increased risks linked to facial recognition and privacy. This project is based in particular on the use of PimEyesa face search engine described by The New York Times as “alarmingly precise”already criticized for its use by the general public. Similar tools like Clearview AI have in the past raised serious concerns about their use by law enforcement. The particularity of I-XRAY is to combine this technology with a consumer device, discreet and easily accessible.
A warning message, not a threat
The students assure that their objective is not to disseminate this tool. “The purpose of this tool is not to encourage misuse, and we will not make it public“, they explain in a presentation document. Their intention is to raise awareness of the risks that this technology already presents today. Thanks to advanced language models, the technology is now able to automate the association of images and personal information from large databases.
The question of confidentiality still under debate
For several years, the issue of privacy in public spaces, particularly with the use of smart glasses, has raised concerns. THE Google Glass had failed largely because of reluctance linked to non-consensual surveillance. However, habits have evolved with the rise of smartphones, vloggers and platforms like TikTok. Today, Ray-Ban glasses Meta are not easily distinguishable from other models, making it difficult for passers-by to discern whether someone is filming them or not. Although the glasses Meta are equipped with a small light that turns on when they are filming, this is often difficult to see, especially in bright or crowded environments. Meta has issued recommendations encouraging users to respect the privacy of others, but this does not guarantee that everyone will follow these rules.
Limited means of protection
This demonstration highlights the potential abuse of this technology, but the two students also emphasize that measures exist to protect one’s privacy. In their explanatory document, they list several databases allowing users to unsubscribe. However, it is almost impossible to completely erase one’s online presence, so the only option left is to make one’s information less accessible.