Meta Ray-Ban glasses are getting a lot of attention at the moment, because Meta has added a dose of artificial intelligence capable of exploiting the cameras outside the glasses. If this is already a big step forward for everyday life, students from Harvard University have gone even further by using glasses for… facial recognition! This allows them to know the identity of each face they encounter on a daily basis.
The students want to show a “preview” of the next step with this type of product
Two Harvard University students, AnhPhu Nguyen and Caine Ardayfio, recently modified Ray-Ban Meta smart glasses using them as support for facial recognition technology, operated through an external program. They demonstrated how connected glasses can scan a face and retrieve personal information online in real time.
Although facial recognition is not directly integrated into the glasses, they capture a video feed which is then broadcast live on Instagram. Intervenes after an external program which is then responsible for analyzing the faces filmed to identify them, recovering information such as name, address or even telephone number, via public databases and online searches. This system uses social networks and specialized platforms to obtain this information, but it is also capable of finding press articles or events in which the person concerned may have appeared.
Nguyen and Ardayfio presented their project by scanning passers-by on the street, accessing their personal information in just seconds, demonstrating the power and speed of the system.
PimEyes and FastPeopleSearch at the heart of the project
Once a face is detected, the external system searches through engines like PimEyes, which matches faces to publicly available images. In addition, services like FastPeopleSearch make it possible to retrieve more detailed information on the identified person. It is mainly these two platforms that do most of the work.
Reactions and issues
Nguyen and Ardayfio assure that their project does not aim to violate privacy, but to show the public the worrying capabilities of technology today accessible to everyone. Meta responded by explaining that the facial recognition technology used by the students, PimEyes, can work with any camera, and is therefore not specifically related to glasses.
PimEyes, for its part, was surprised by this use and recalled that its tool was not designed to allow personal identification, either directly or indirectly.
Are we ready for a world where our data is exposed at a glance? @CaineArdayfio and I offer an answer to protect yourself here: pic.twitter.com/Oo35TxBNtD
— AnhPhu Nguyen (@AnhPhuNguyen1) September 30, 2024
Implications and ethical questions
This experiment vividly demonstrates how it is possible to collect personal information about a person without ever having spoken to them, or even knowing their name beforehand. If two students with limited means were able to achieve such a feat, it is worrying to think what companies with greater financial means could achieve.
In France, such use of facial recognition would be strictly illegal in public spaces. GDPR (General Data Protection Regulation) prohibits an individual from using this technology in public places, in order to protect the privacy of individuals. The legal framework imposes strict restrictions to avoid any abuse.
Faced with this technological advancement, Meta has not expressed any particular concern about the possibility that other Ray-Ban Meta users could exploit these facial recognition capabilities for similar purposes and no measures have been announced to supervise these practices.
