Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
AI Crime

New Orleans OKs Some Police Use of Facial Recognition (apnews.com) 34

An anonymous reader quotes a report from The Associated Press: The New Orleans City Council has reversed itself and approved police use of facial recognition software and cellphone surveillance towers to investigate violent crimes. The ordinance, approved by the council on a 4-to-2 vote Thursday, comes as killings in the city reach numbers last seen in the mid-2000s after Hurricane Katrina. It partly reverses an ordinance passed nearly two years ago, when crime was low. Mayor LaToya Cantrell called it "a tremendous stride towards greater public safety."

The ordinance lists 39 specific crimes that can be investigated by using the technologies, including murder, rape, stalking, and battery of a police officer. Two other kinds of policing software remain forbidden: programs that seek to predict spots where crime is likely and those which use characteristics such as size, clothing or vehicle model to track people.
The American Civil Liberties Union of Louisiana said facial recognition software has been shown to be biased by race and sex, and "there is absolutely no evidence that reinstating facial recognition will help reduce violence."

Council member Eugene Green, who proposed the ordinance, said new police policies, including procedures for ensuring accuracy, were adequate safeguards. John Thomas, director of public safety and homeland security for the city, added: "The facial recognition in and of itself cannot get you any arrest warrants, no search warrants. It is just a tool for us to say, 'OK, this is a lead.'"
This discussion has been archived. No new comments can be posted.

New Orleans OKs Some Police Use of Facial Recognition

Comments Filter:
  • Maybe the cameras should focus on detecting crime [theadvocate.com] instead of uselessly tracking people?

    • Maybe the cameras should focus on detecting crime [theadvocate.com] instead of uselessly tracking people?

      How did you get that from your linked article? To me the complaint seems to be that there is an increase in crime committed with stolen guns. Often stolen from cars. some cars being unlocked and thefts not reported because people don't write down the serial numbers of their guns.

      • because people don't write down the serial numbers of their guns.

        Y'know, I've owned guns for nearly 50 years. Got to admit it NEVER occurred to me note the serial numbers of my guns...

        • Y'know, I've owned guns for nearly 50 years. Got to admit it NEVER occurred to me note the serial numbers of my guns...

          It would seem like a good idea to at least take a pic for insurance purposes.

      • How did I search the internet to see what the most current problems in New Orleans are? Or how do I make the connection that a camera system that can detect objects could exist? I had hope the link that a city is complaining about an increase of crime because of X should lead to looking for more of X. Instead the same city wants a face tagging system that appears unrelated to the actual crime they are actually complaining about.

        If people are walking around with a (potentially stolen) gun, the odds are slig

  • Inadequate oversight (Score:4, Informative)

    by drinkypoo ( 153816 ) <drink@hyperlogos.org> on Friday July 29, 2022 @08:12AM (#62743988) Homepage Journal

    As we've seen, even requiring a judge to sign off is inadequate oversight. But this ordinance apparently doesn't require even that.

    Council members J.P. Morrell and Lesli Harris said the ordinance wonâ(TM)t improve public safety and will divert focus from other urgent problems in the police department.

    [...]

    Morrell and Harris said they could not vote for the ordinance without changes to ensure that it couldnâ(TM)t be used against same-sex couples and people seeking abortions, news agencies reported. Their proposed amendment also would have required a judge to sign off on use of the technology and regular reporting on its efficacy.

    It doesn't have even the level of oversight implied by the warrant process (which is itself next to nothing) and it doesn't require anyone to prove that it's useful. Pucker factor ten.

  • “The facial recognition in and of itself cannot get you any arrest warrants, no search warrants.”

  • The American Civil Liberties Union of Louisiana said facial recognition software has been shown to be biased by race and sex

    In which way? We are talking a software that recognizes faces. I mean it recognizes better some races, or worse? And which option is better, I mean, less biased? If black women are recognized more precisely is that bad because they have less privacy or good because they have to fight less false positives?

    I'm sorry but I fail to see how such software can be biased, it can be less precise in some instances, I'm sure, but then it just has to be improved. You can be falsely recognized as easily by a human offic

    • The sensors get poor quality data when there is not a lot of contrast due to dark skin tones. Light skin tones reflect more light, that's just physics.
      Feed poor data to the software, which doesn't try very hard to qualify the data it is given, and you get poor quality results.
      Q.E.D. garbage-in / garbage-out.

  • Face recognition is a great technology, at least when done right. Police and politicians are low-intelligence actors who do not understand its limitations, or when they are being bullshitted by vendors. Activists are evil people who ascribe racism to anything that harms their favorite group. Bumpy road ahead.

  • WTF is wrong with: 'programs that seek to predict spots where crime is likely'?

    I was working on stuff like that over 20 years ago just mapping where crimes happened - or at least reported crimes happened. We used first generation Pentium PCs; we're not talking 'AI' here.

    Or does that sort of information reveal information which is problematic for community relations? Or indeed prove that the police spend too long patrolling wealthy areas where nothing happens.

    • WTF is wrong with: 'programs that seek to predict spots where crime is likely'?

      You don't need a program to do that. Any cop in any city can tell you where more crimes are likely to occur. Simple day-to-day experience will make it obvious.

      When you increase enforcement in areas that have more major crime, you also (as a consequence of just having more cops looking for crimes in the area) catch more people committing minor crimes. This leads to disproportionate enforcement of minor crimes in various areas.

      This is good in theory: people learn that committing crimes in this area will be

      • Not convinced. Hot spot analysis - at least as we did it - was concerned with real offences that disrupt the lives of the victims: burglary, theft of / from vehicles, muggings. The iffy ones you might question were assaults and criminal damage. But even those can have substantial effects on the victims. The assumption that 'people in poor or minority areas are punished for trivial offenses' needs to be demonstrated, whilst the claim 'people in wealthy white neighborhoods are allowed to break laws at will' o

        • The problem is that the police should be neutral and spend their resources evenly when there is no active crime being committed. I.e. There shouldn't be more cops in one area than the other unless they are responding to a complaint / crime in progress. Doing otherwise is just pandering to the people in charge and feeding their political narratives.

          The assumption that 'people in poor or minority areas are punished for trivial offenses' needs to be demonstrated

          Neighborhood A has 10 cops.
          Neighborhood B has 30 cops.

          If the crime rate of both neighborhoods are equal, which neighborhood has a higher chance of a crime bei

          • What is the possible point in allocating police to areas where no crime has been recorded in decades?

            and it's not about police 'witnessing' crimes, it's about police responding to crimes that are occurring - catching the burglar in the premises, chasing down the mugger or car thief. To do that they need to be where those crimes are occurring. And, as a bonus, they may deter them from occurring.

            • Humbug

              Well that's nice. You consider my comment to have no merit on it's face. This will end well.

              it's not about police 'witnessing' crimes, it's about police responding to crimes that are occurring

              Nice deflection. At what point did I say that police shouldn't be responding to a crime in progress? Answer: I didn't.

              Now answer the question: Why did you say "The assumption that 'people in poor or minority areas are punished for trivial offenses' needs to be demonstrated"? When a basic third grade math problem will demonstrate it for you. Be specific.

              What is the possible point in allocating police to areas where no crime has been recorded in decades?

              they may deter them from occurring.

              According to you, to deter crime.

              Now answer this question:

In the long run, every program becomes rococco, and then rubble. -- Alan Perlis

Working...