LONDON — Intelligence agencies, online advertising companies and hackers can still figure out which YouTube videos a person watches despite the site’s efforts to keep users anonymous, according to a new study from Ben Gurion University of the Negev.
“We built a simple yet robust machine-learning algorithm that can identify which video you watched — within a predetermined set of videos — with a high degree of accuracy,” says Dubin in a university news release. “The algorithm is based on an in-depth study of how video services work, how video content is encoded and how a video player requests information to play it.”
Dubin used this algorithm to find out if someone had watched specific videos from a series pertaining to terrorism and hopes his technology could help track potential terrorists.
Though Dubin thinks his algorithm will be helpful, he warned everyday YouTube users that what they watch and search for can be traced — even by marketing companies looking to record consumers’ activity.
“It’s important to know that video encryption is not as secure as we once thought,” he says. “Google, YouTube’s parent company, is not likely to patch the gaps, since it would be prohibitively expensive to create a traffic obfuscation mechanism for every user’s every video request.”