Predictably low trust by Black Americans in facial recognition in new poll

Every answer in a new Pew Research Center survey about police use of facial recognition makes it clear that Black Americans expect more to be targeted than helped by the software.

Most media interest in this news will focus on (as does this article) racial disparities. While that is a critical result to understand, so is the aggregate view of police use of facial recognition.

Barely a third of U.S. adults as a whole think policing would be improved just by putting the technology in neighborhoods.

The only reasonably clear hope that facial recognition could be a positive addition to policing is if departments train officers in how systems can make errors, and even then, it would only make their use “more acceptable.”

Views are role-dependent, too. More than half (sometimes far more) of respondents said it was acceptable to perform face scans of crowds entering large events or those at public protests.

Among Blacks, whites and Hispanics, only Blacks uniformly reported that they would expect to be needlessly monitored by and arrested because of facial recognition.

Levels of trust vary by racial group and the technology’s specific roles, but Black people expressed the most negative views.

That is not to say all white and Hispanic people see facial recognition as a net benefit for them and their communities – far from it. A quarter of all those taking Pew’s survey last November said widespread use of facial recognition would make policing less fair.

Breaking that question down, just 34 percent of all respondents said they think widespread use of machine vision would make policing fairer. Forty percent said it would make no difference.

Hispanic survey takers were the most optimistic, but still, only 40 percent think policing would be fairer. Thirty-five percent felt facial recognition would make no difference.

Of white respondents, just 36 percent said they would expect fairer policing because of AI algorithms. A like percentage felt nothing would change.

Blacks answering the survey were the most distrustful. Only 22 percent would expect fairer treatment and 47 percent said facial recognition would make any difference.

Half of Black survey takers said they are certain that Black and Hispanic neighborhoods would be monitored. Only 18 percent of Hispanics agreed with that statement and whites were in the middle.

More false arrests of all groups will “definitely” occur, in the view of Black respondents.

False arrests of Black Americans based on facial recognition have occurred, in violation of police procedures, but even determining the veracity of some of the few claimed incidents can be challenging.

A better survey might ask taxpayers how much should be spent in developing trust. There are some answers already, but the topic still lacks focus.

Article Topics

accuracy  |  biometric-bias  |  biometrics  |  facial recognition  |  market research  |  police  |  surveillance