An In-Depth Analysis of the Privacy Risks of Photo Metadata
Modern digital photos contain EXIF metadata that can expose up to 30 categories of technical parameters, the most privacy-sensitive of which are geographic coordinates and time-series information. Tests have shown that 78% fully records the latitude and longitude coordinates of the location in photos taken by GPS-enabled smartphones (with an accuracy of up to within 5 meters), which can be pinpointed to specific buildings when combined with Google Maps data. Meanwhile, 100%'s digital photos all contain precise time stamps (accurate to the millisecond level) of the shooting, and these timing data may expose the user's life pattern.
They even contain device fingerprint information such as camera serial numbers, making it possible to correlate photos uploaded from different platforms for traceability. In contrast, the visual content recognized by AI is more of a probabilistic inference, while metadata belongs to deterministic information. Professional tests have shown that by analyzing the metadata of 50 social media photos, it is possible to reconstruct the trajectory of a user's 90% daily activities. This is exactly the warning purpose of Ente team to develop this tool - to make ordinary users aware of the risk of invisible data leakage.
This answer comes from the articleThey See Your Photos: Analyzing Photo Privacy Information Based on Google VisionThe
