Stanford Students Develop AI That Can Identify Photo Locations, Sparking Privacy Concerns

Despite its potential applications, the technology raises serious privacy issues and calls for regulation and responsible development.

Three Stanford graduate students have developed an AI model, called Predicting Image Geolocations (PIGEON), that can identify the location of a photograph with 95% accuracy and within 25 miles of the real location.
The AI model was trained on Google Street View data and can have real-world applications such as identifying roads with downed power lines or assisting in biological surveys of regions.
Despite its potential positive uses, the technology has raised serious privacy concerns. Jay Stanley of the American Civil Liberties Union warns of potential misuse in government surveillance, corporate tracking, and stalking.
The student creators, aware of these concerns, have decided against releasing the technology to the wider public.
Regulation is needed to prevent abuses of this technology, and companies developing AI tech must work to prevent damage caused by their products.