Millions of Online Photos Used Without Consent to Build Facial Recognition AI

People’s faces are being used without their permission to power technology that will eventually be used to surveil them. The facial recognition technology is based on feeding millions of images of faces into algorithms to help the systems determine the hundreds of ways each face is unique.

The photos are coming from the internet, where they’re swept up by the millions without the knowledge or consent of the people who posted them, categorized by age, gender, skin tone among dozens of other metrics. The data is then used by corporations to build algorithms to track and identify people.

IBM is the latest company to enter the arena, the company released a collection of around one million photos that were pulled from Flickr and processed to describe the subjects’ appearance. IBM’s entrance into the facial recognition surveillance technology is worrisome as the company was previously sued by concentration camp survivors for the work to assist Nazi’s in tracking and identifying people during WWII.



“Diversity in Faces(DiF)” by IBM Research is licensed under CC BY-ND 2.0