Our household bulletin board, the one we call Facebook built a system that collects users’ personal data, such as education, homeownership and internet usage, in order to predict their socioeconomic status. This is not necessarily new or interesting, but what it is, that the AI then classifies people as belonging to one of 3 pre-determined groups: working class, middle class and high flyers.

While the initial dataset is said to be based on a voluntary questionnaire, it could work without any direct user input, because it could fetch data by filtering the stored data of past user interactions. The questionnaire is a decision tree, which is funny, as this is also the basis of most dataset-fed machine-learned AIs. In the patent’s text it is described in this way: “not only uses data supplied by users on the platform, but can also refer to actions performed by the user on [Facebook]”.

Day in, day out there is a newer way to surveil and analyze people in a more and more refined way. There is a growing tendency to see the world in binary good/evil and poor/rich terms and patents like this will not help to put an end on the black-and-white outlook on life of our betters.

Sources:

https://www.freshpatents.com/-dt20180201ptan20180032883.php

https://www.dailymail.co.uk/sciencetech/article-5346733/Facebook-patent-uses-track-socioeconomic-status.html