Which photos on social media have become dangerous due to AI development

The article explains how AI systems can extract geolocation data from photographs, even after EXIF metadata has been removed. Neural networks analyze architecture, signs, nature, and vehicles in images to determine location. This information can be used by scammers for personalized attacks. Recommendations for protection include checking the background, cropping, and privacy settings.

The article warns that posting photos on social media can reveal much more information than it seems. Modern AI systems, such as ChatGPT, Gemini, and Claude, are capable of analyzing not only obvious landmarks but also subtle details in the background to determine location. Removing EXIF data is no longer a sufficient protective measure, as algorithms extract information directly from the image. Neural networks analyze architecture, signs, nature, cars, and road infrastructure. In tests, AI was able to identify a street in London based on buildings and traffic, as well as a country from a photo of a forest road. This information can be used by scammers to create convincing phishing schemes and track routes. The article provides practical advice: check the background, blur recognizable objects, change cropping, adjust social media privacy settings, and use secure messengers. The material is based on a publication by Popular Science.