In brief: Facebook is always looking to improve the efficiency of its data mining practices, and it may have found a way to do just that if a recent patent is anything to go by. The patent describes an algorithm that could examine your family photos to better target you with ads.
In Facebook’s own words, the system would work as follows:
An online system predicts household features of a user, e.g., household size and demographic composition, based on image data of the user, e.g., [profile photos], and textual data in the user’s profile that suggests relationships among individuals shown in the image data of the user… The online system applies one or more models trained using deep learning techniques to generate the predictions.
Obviously, algorithm-based advertisements are far from new, and they aren’t necessarily nefarious. Even the staunchest privacy advocates generally seem to accept that these practices are the norm.
However, there may be a good reason for said individuals to feel particularly uneasy about this patent. After all, Facebook’s proposed system may allow advertisers to obtain information about your family without you explicitly surrendering that information to the company.
Facebook’s proposed system may allow advertisers to obtain information about your family without you explicitly surrendering that information to the company.
For example, even if you don’t list your marital status on the platform and your family members don’t use Facebook at all, this algorithm might still be able to detect that you have, say, three children and a wife purely by looking at the images you’ve uploaded. It could use this information to target you with ads geared towards your family composition.
If you have three daughters, perhaps the algorithm will suggest feminine products geared towards kids, even if they aren’t products you’d buy for yourself.
Regardless, it should be noted that this is indeed just a patent at the moment, and not necessarily an indication of Facebook’s future plans. Tech companies patent new ideas (many of which are even more outlandish than this) all the time, and very few of them have been announced, much less become a reality.