NYT Facial Recognition Experiment

An experiment reported in this weekend’s New York Times by graphics editor Sahil Chinoy caught my attention for its resourcefulness and implications.

The experiment involved (1) scraping images of faces of people visiting Bryant Park from web video captured by cameras mounted near the NYPL (2) running the images through through low-cost facial recognition technology (3) identifying individuals who work in the neighborhood by matching the scrapes with head shots from their employers' web sites, and (4) trying to contact them to talk about the implications of facial recognition technology and privacy. 

First the numbers … 

9 hours of footage from cameras in Bryant Park

2,750 faces detected 

$60 cost to run FR service

1 match

0 federal, state or city laws broken

Interesting, but that’s not the best of it. 

The Times contacted the match to confirm his identify and discuss the experiment with him.

It was an effective/affective way to have a conversation about the implications of FR technology, and how to think about its impact on expectations of privacy.

With +9,000 cameras placed systematically throughout lower Manhattan (many of them security cameras built into wifi-providing LinkNYC kiosks) and 0 federal, state, or city laws restricting the use of FR … the possibilities are hugely interesting and somewhat concerning. 

If one of the purposes of doing the experiment was to raise awareness of this situation, it worked for me. 

You can see the interactive feature here …

https://www.nytimes.com/interactive/2019/04/16/opinion/facial-recognition-new-york-city.html