AI can identify photo locations to fuel phishing scams

Research from McAfee reveals that artificial intelligence can accurately identify the locations of travel photos posted on social media platforms like Instagram and Facebook, even without GPS metadata or location tags. In a study of over 21,000 images, two AI models achieved geolocation accuracy rates of 91% and 87%.

Scammers can exploit this capability to conduct highly convincing phishing attacks. By analyzing visual cues such as landmarks, street signs, architecture, and even specific vegetation or lighting, AI can pinpoint a user's location. Criminals can then use this information to send fraudulent messages—such as fake bank alerts regarding unusual activity in a specific city—to trick victims into revealing sensitive financial data.

Experts suggest that even seemingly generic images, such as a hotel room or a river scene, can provide enough context for AI to identify a country or a specific region. To mitigate these risks, users are advised to avoid posting photos in real-time, limit the visibility of their social media posts, and carefully check for identifiable landmarks before sharing images.