Facial recognition is becoming more and more common these days but it is also one of the more controversial forms of biometrics because of its privacy implications. There is, however, one other problem with face ID that rarely pops up in reports: racial bias. Some users of face recognition technologies report that these systems seem to be designed to work better on people with fairer complexions. Given its prominence, Google is working to ensure it won’t be accused of racial bias. Reports claim, however, that it is doing rather questionable methods, including profiling, to accomplish that.
It was already reported before that Google took to the streets to take pictures of complete strangers in exchange for monetary compensation or Google Play coupons. The goal is to feed its machine learning with enough data to improve face recognition. While the voluntary method isn’t exactly illegal, some found it distasteful on a different level.
This time, however, the allegations are more serious. According to the Daily News, Google sent out workers to record people with questionable methods. Those include targeting people of color specifically with the goal to teach its neural networks against racial bias, a rather ironic method to get to an otherwise laudable goal.
It doesn’t stop there, however, as the tech giant has allegedly gone beyond paying people to give up their faces. Methods include not telling people they are being recorded, which will violate privacy laws, to denying that fact when discovered. Some were also told to only approach homeless people because they are less likely to refuse or report incidents.
Such exposés are not uncommon in the tech world, especially when it comes to tech giants like Google. The Daily News offers numerous anecdotes and testimonies of Google employees discretely going about the facial data collection with little to no knowledge of people. Google will undoubtedly refute the report but the red flags raised could initiate an inquiry into these alleged methods.