Though Facebook filed the patent on May 10, an application published Thursday detailed an algorithm that can use family photos to analyze faces and other elements, cross-reference them with the rest of you and your family’s online personas, and figure out every member in your household.
Using data mining, the proposed algorithm would identify a posted photo’s identifying tags and descriptions along with a user’s IP address and track down other Facebook users with matching information, thus determining a household’s size, makeup, and interests—even who has a phone and who doesn’t. What exact demographic composition it could identify wasn’t specified in the application, and the patent does not extend to video or any content conveyed in private messages.
This system would work in tandem with the family targeting option Facebook first enabled last year, which uses location data among other information to better target people living together, as well as its Portal video chat device with face-tracking technology.
- Facebook is targeting ‘borderline content,’ creating independent appeals group
- Facebook reportedly hired PR firm to write critical articles about Apple, Google
- Facebook patches vulnerability that could have exposed user data
H/T the Verge