IPVM has repeatedly investigated Facewatch, the UK company selling live facial recognition to retailers to catch shoplifters, e.g. Facewatch's Hikvision usage, poor signage, and GDPR claims.
Now, The New York Times has published a feature on the company (Barred From Grocery Stores by Facial Recognition) saying it is "inside nearly 400 stores" in the UK by licensing face recognition software from Real Networks and Amazon:
*********, ***** ******** ****** *********** ******** made ** **** ******** *** ******, is *** ****** ****** *** ****** across *******. ******* ** ******** ** pictures *** ******, *** ******* **** the ********* *********** ** * **** as *** ****** ***** **** * shop *** ***** ** ******* * database ** ******* ******
******* "******** ******", **** ** * woman *** **** *** *** *** was *********** ***** ** *********'* ****** after ************ (********* ** ***) *********** about $** ** ***********:
***** *** ********** *** ********* ********** the *****, ** *** *** ***** much **** *** ***** **********. ******* Facewatch *** *** ***** ***** *** incident ******** ********* *** ** *** her **** **** *** *** ** the ********* *** ** *** **** had ********.
*** ***** **** *** *** *** recall *** ******** *** *** ***** shoplifted. *** **** *** *** **** walked *** ***** *** ********* **** her ***** **** ******* ****** ** go ******* ** * ****-******** *****.
********* ** ***** ***** ** *** UK ***** *** *******'* **** ***************** *** ********* ** ***** ****. *******, ** **** **** **** limits, *.*. ******* ****** ** ***** automated *** **** *** * ***** "super **********" ** ******* **** ***:
***** **** *********’* ****** ********** * shoplifter, * ************ **** ** * person *** ****** * **** ** be *“***** **********”— ******* **** * ******* ****** for *********** *****. ****** *******, *** super ********** **** ******* *** ***** against *** ********* ******** ****** ** alert ** ****.
**** "***** **********" ************** ******** ** "******** ******" ** ****** Carlo, ******** ** ******* ****** ***** Big ******* *****, ********* ********* ******* *************** ** **** *********, *** *** limit **** ** *** ********* **** as ********* **** "******" ********* *** be ***********.
********? **** ************* *** ****, ***** *** ******** ********* (*** all *** *******/***** ****** ** ***** bring)?
******** "*********" ******. **** ** ** this ****. *** ********* ********** ***** according ** *** *****, *** *** allegedly ****** ** *** ***, *** didn't ****** ***. ****** *** ** something ***** ***** **** ********* ** common **** ***** *** ****'*- **** both ***** **** ****'* ** ******** wrong. **** ** *** ** ***** many **** ***** **'* **** ** get **** ******* ** **** ** more ***** ** **** ***.
** **** **** * ****** **** to *** ******* **** * ***** for ********* ******** ** **** $** a **** ******** ***** ******** ** happen. ** ** ***** ** ** believe ***** ** ******** **** ** the ***** **** ** ****.
***'* **** *** **** ***** **** done ****. *** ******* ***** *'** gone ** *** ***** *** ** idea ***** * **** *** *** my ******* ****. **** ** **** had * ****** ** * ******** transaction, ***** **** ******* ** ******* info ******* **** ****** ** ** through *** ****/****** **** *******?
*** ***'* ******* *****Barred **** ******* ****** ** ****** *********** is somewhat disingenuous. She wasn't barred by a system working solely on facial recognition. She was barred by the store using facial recognition as a tool, not a whole lot different than printed pictures of people who passed bad checks or shoplifted before that I used to see at grocery store checkouts. Facial recognition is just being used as a more efficient version of that system. But it was still a person reviewing the stored evidence and the scanning results who made the decision she matched and therefore should be barred. (As far as I could tell. I couldn't read the article behind a paywall.)
************ **** **** ***** *** ******* to ******* **** ** *********, ****'** not ******* ********** **** ****** *** to *** ******** **** *** ** stop *** ******.