81% of 'suspects' flagged by Met's police facial recognition technology innocent

Re:81% of 'suspects' flagged by Met's police facial recognit... by kurthr on Friday July 05, 2019 @03:56PM (#20365355)

For each individual crime there probably is only one or a small number of perpetrators. You can't just say... hey we found 1000 (or even 5) of you face matched for a single crime (assault) which you almost certainly (or even probably) didn't commit, but after a search or warrant some of you appear to have committed a different other crime (drug possession, tax evasion, public urination, evading arrest, immigration status) so go to jail and wait for court.

Once police have the right to search, enter, confiscate. and arrest probable cause is out the window.

[ Parent | on HN ]

Re:81% of 'suspects' flagged by Met's police facial recognit... by educationdata on Friday July 05, 2019 @04:36PM (#20365666)

I don't follow. Do they match faces to crimes or match faces to faces of criminals?

[ Parent | on HN ]

Re:81% of 'suspects' flagged by Met's police facial recognit... by kurthr on Friday July 05, 2019 @06:28PM (#20366387)

They match faces out of a crowd for criminals, but to be a criminal you have to have committed a specific crime... usually that means a single person. For any one crime, they are looking for a single person and that's worst case for false positive recognition. For a large sample set 50k-1000k, they are lucky they don't match any one person to multiple criminals (birthday paradox) ... that would be obviously disqualifying.

I suppose it could be worse and they just take an aggregate face of all (accused?) criminals or for particular crimes (white collar?) and match against that... hey your skin was brown, or hey you didn't shave... you appear Male.

[ Parent | on HN ]