This post is also available in:
עברית (Hebrew)
A group of Harvard students developed innovative AI glasses that allow the user to instantly know someone’s personal details, such as their name, address, and phone number, just by looking at them. However, their aim is not to unleash this technology on the public; instead, they seek to raise awareness about the potential risks associated with advanced facial recognition and data extraction technologies.
The glasses, dubbed I-XRAY, combine several cutting-edge technologies that have only been released in recent years. Using Meta Ray-Ban 2 smart glasses, the system captures images of individuals in public spaces. The captured image is then processed using facial recognition models similar to those employed by popular search engines like PimEyes and FaceCheck.id. These models analyze the face and compare it to publicly available images online, returning relevant URLs associated with the individual.
The students reported that PimEyes proved to be particularly effective in their tests, accurately identifying a diverse range of individuals. Once the relevant URLs are gathered, an advanced Large Language Model (LLM) extracts detailed personal information—such as names and occupations—from the unstructured data available online.
The capabilities of I-XRAY bring to light significant privacy concerns. The tool leverages resources like FastPeopleSearch, a service that compiles publicly available information into searchable profiles. Users can obtain details such as home addresses, phone numbers, and even relatives using just a person’s name. This raises troubling questions about how easily sensitive information can be accessed.
Furthermore, the use of platforms like Cloaked.com, which trades phone numbers for partial social security numbers, illustrates how deeply interconnected and accessible personal data has become. The students emphasized that once I-XRAY identifies a person’s name, a quick lookup on FastPeopleSearch could reveal a wealth of personal details, including home addresses and phone numbers.
While the I-XRAY project showcases the potential of AI in facial recognition and data extraction, it serves as a cautionary tale about the implications of such technology. The students are not planning to release I-XRAY to the public, stating, “Our goal is to demonstrate the current capabilities… to raise awareness that extracting someone’s (personal details) from just their face is possible today.”
Fortunately, there are ways to protect your privacy. Services like Pimeyes and FastPeopleSearch allow individuals to request the removal of their photos and personal data. By highlighting the current capabilities of smart glasses and the ease of accessing sensitive information, I-XRAY underscores the urgent need for discussions about privacy rights in our increasingly digital world.
