AI Pinpoints Your Location with 91% Accuracy

I just returned from a week spent in a certain country, and I posted some photos that did not feature prominent landmarks from the trip I was taking, without adding a geographic location or writing the city’s name in the caption—just a nice picture. A few days later, I received a message pointing to the bank I deal with, noting suspicious activity “during your trip abroad.” The message was disturbingly specific, with details about where I was and when I was there, and it appeared genuine. This type of targeted phishing scam has become increasingly common, and your photos themselves may have helped craft it, according to McAfee, a cybersecurity and computer protection software company. A team at the company sought to determine how much geolocation information can be extracted from an ordinary travel photo, and what that means for the millions of people who travel each year. The team concluded that the findings require a change in how you think about what you share online. They found that some artificial intelligence models achieve an accuracy rate exceeding 90% in identifying where a photo was taken, based solely on the visual elements within it. More importantly, this level of accuracy is now available using free, widely accessible tools. According to Al Arabiya, the question the team wanted to answer seemed straightforward on the surface: Can artificial intelligence look at a travel photo and determine where it was taken, even in the absence of GPS data or geotags? This is not about metadata or embedded coordinates, but about the image itself: the background, architecture, signs, lighting—essentially, the visual context that any photo naturally captures.