Live data from Hacker News

Santa Cruz, California bans predictive policing in U.S. first

reuters.com

161–170 of 643 posts

Re: Santa Cruz, California bans predictive policing in U.S. first

#161

Looking at all these comments, did no one watch Minority Report ? The whole film was about the difficulties and ethics of "precrime" and arresting people before they even committed a crime. Obviously we aren't doing that here, but stories often exaggerate a bit. I'd also like to add another good series that explores this topic even more than Minority Report : Psycho-Pass . Psycho-Pass is more extreme in that there's…

Big budget movies sort of struggle against telling stories about justice for people whom we discriminate against contemporaneously. You could never make a movie that even loosely portrays the plight of Uighers, who suffer under a notoriously technology-heavy form of surveillance and oppression (what we are calling predictive policing). Not because of censorship but because no one is going to fund your film. So while…

I think the purpose of movies and stories is to make these events relatable. You're supposed to draw parallels. You don't have to show a black person to convince white people that minorities are being repressed. In fact, it is probably more convincing if on screen you are showing white people being oppressed because then they are going to more identify with that character. Of course, we are supposed to analyze these stories and find what they are trying to tell us. They aren't always so much "here's the entire problem, why you should care, and every little nuance" but rather "this is a problem." It often serves as a stepping stone for people to learn and start to care. To see things that they didn't see before because a story can give you a glimpse into what it is like living in someone's shoes. I wouldn't bash a story for not being able to compact what requires dozens of hours of research into an entertain-able bite size chunk. Stories serve different purposes from documentaries which serve different purposes from literature and news.

Re: Santa Cruz, California bans predictive policing in U.S. first

#162

> Used by police across the United States for almost a decade, predictive policing relies on algorithms to interpret police records, analyzing arrest or parole data to send officers to target chronic offenders, or identifying places where crime may occur. The promise of the software made in a sales pitch or on the website, is a far cry from the reality of what such software delivers. That's fine when it's business, t…

Predictive policing does nothing more than sending policing resources where crime is more likely based on statistical models. It does not racially discriminate and it is not racist. It does not harm anyone. I feel that this is posturing and shooting the messenger. If crime is statistically higher in "black neighbourhoods" the issue will not be solved by pretending it isn't. If the technology does not work then of cou…

Everyone does something criminal from time to time. Policing A more than B will lead to more arrests at place A, increasing their crime statistics, leading to more policing ...

For example drug use is about equal between white and black Americans, but since black people are more strictly policed (more frequent in traffic stops etc) they are arrested and sentenced for drug use far more often than white Americans. [0]

0: https://www.hamiltonproject.org/charts/rates_of_drug_use_and...

Re: Santa Cruz, California bans predictive policing in U.S. first

#163

Earlier quoted context omitted.

> This is a premature application of software. maybe, but you have no idea yourself so your fear does not carry enough weight to carry that statement. It's basically like you are saying "it's too early to try Bayesian policing" when in fact Bayesian policing might be a great idea, and it was a great idea 10 years ago too. Conditional probability, if there have been a lot of rapes in a 2 block radius, do you think it…

Why not take your Bayesian decision theory and make a trading system? Invest your farm, and you'll make millions with ideas like, if a stock just had two upticks, isn't it a good idea to put a little money into an upward trending stock . Not only will you come to understand decision theory better, behavioral prediction better, you'll get paid to do it, and nobody else will get hurt!

The stock market isn't the same thing...

Re: Santa Cruz, California bans predictive policing in U.S. first

#164

Earlier quoted context omitted.

The median household income in Middlesex is over $100K (that's where Harvard is located, right next to MIT). Incarcerations were also flat in Martha's Vineyard, but so what? My priors say Middlesex wouldn't show increased rates of incarceration regardless of the amount of policing, the passage of three-strikes laws, the passage of the Clinton crime bill, etc. due to the demographics of the area. What do your prior's…

And yet, the crime rate in that county (and in Cambridge and Lowell) reflects the same plunging nature as the rest of the state and nation, since 1990 peak. Take the big CSV from the github I posted, combine it with the FBI UCR data, and try to find a correlation between incarceration rate and crime rate.

I have, it's the main topic I study. I've just never met anyone with the view that increased incarcerations rates are not associated with decreases in crime (specifically, violent crime), and I don't think the data supports that either. I literally have no idea how you came to that conclusion given the stats available to researchers over the last 40 years.

I'm not suggesting that there aren't other ways to decrease crime besides incarceration, there are many possibilities. But I don't think there's any data showing that incarcerating criminals doesn't reduce (violent) crime.

So I'll just assume we agree in general, and that our definitions are different somehow, since we seem to be looking at the same datasets.

Re: Santa Cruz, California bans predictive policing in U.S. first

#165

> Used by police across the United States for almost a decade, predictive policing relies on algorithms to interpret police records, analyzing arrest or parole data to send officers to target chronic offenders, or identifying places where crime may occur. The promise of the software made in a sales pitch or on the website, is a far cry from the reality of what such software delivers. That's fine when it's business, t…

I have no problem with that, but why aren't these types of statistical models and curve analysis being used to deploy drive by patrols? Some of this seems like throwing out the baby with the bath water, does it not? While I agree that it's way too soon for this sort of granularity, is it too soon to have 10 block, increased presence of troubled areas helped by statistical analysis?

> I have no problem with that, but why aren't these types of statistical models and curve analysis being used to deploy drive by patrols? Some of this seems like throwing out the baby with the bath water, does it not?

Combined with the evidence of extreme LEO abuse of power, doesn't this reinforce the idea that certain people/neighborhoods become disproportionately affected? Isn't this exactly what the past few weeks of protests have been railing against?

Re: Santa Cruz, California bans predictive policing in U.S. first

#166

> Used by police across the United States for almost a decade, predictive policing relies on algorithms to interpret police records, analyzing arrest or parole data to send officers to target chronic offenders, or identifying places where crime may occur. The promise of the software made in a sales pitch or on the website, is a far cry from the reality of what such software delivers. That's fine when it's business, t…

Predictive policing does nothing more than sending policing resources where crime is more likely based on statistical models. It does not racially discriminate and it is not racist. It does not harm anyone. I feel that this is posturing and shooting the messenger. If crime is statistically higher in "black neighbourhoods" the issue will not be solved by pretending it isn't. If the technology does not work then of cou…

> where crime is more likely

This is the fundamental disconnect in your argument, because historical recorded arrest rates are not necessarily reflective of actual crime rates.

Re: Santa Cruz, California bans predictive policing in U.S. first

#167

Earlier quoted context omitted.

> Obviously we aren't doing that here IDK. one year olds were added to CalGang database. People were added to the CalGang with no supporting evidence. The database was used to direct police resources and also for background checks for job applicants . So, you're entered into the database as a one year old (or a ten year old, or a 30 year old but for no good reason). Now you can't get a job. Predictably, you turn to w…

I meant specifically arresting people before they committed a crime. There is a fine line between that and putting someone under watch, but I think Minority Report was pointing out that this line might as well not exist. Personally, I think putting someone under watch before they have committed a crime or there is reasonable suspicion that the person is about to commit a crime is unconstitutional. I do not think stat…

It's not just putting them under watch. It's putting them under watch while also using the same database to screen job applicants.

Pushing someone out of the legal economy and watching closely has an almost guaranteed outcome. Doing those two things simultaneously is functionally equivalent to either a) starving them to death or b) precrime.

Re: Santa Cruz, California bans predictive policing in U.S. first

#168

Earlier quoted context omitted.

> This is a premature application of software. maybe, but you have no idea yourself so your fear does not carry enough weight to carry that statement. It's basically like you are saying "it's too early to try Bayesian policing" when in fact Bayesian policing might be a great idea, and it was a great idea 10 years ago too. Conditional probability, if there have been a lot of rapes in a 2 block radius, do you think it…

Yea but you're ignoring biases in the data that are difficult or impossible to disentangle from the underlying effects. What happens when you see an area with 50% higher violent crime rate? Is that because there are a large number of crimes? Or is it because it's a neighborhood that is disproportionately targeted by police? How do you even attempt to control for that? It's not that Bayesian stats can't help or can't…

Our knowledge of crime rates doesn't come form arrest records, but rather victims surveys. Yes, we know that certain neighborhoods/populations have higher rates of crime and not just arrests.

Re: Santa Cruz, California bans predictive policing in U.S. first

#169

> Used by police across the United States for almost a decade, predictive policing relies on algorithms to interpret police records, analyzing arrest or parole data to send officers to target chronic offenders, or identifying places where crime may occur. The promise of the software made in a sales pitch or on the website, is a far cry from the reality of what such software delivers. That's fine when it's business, t…

I have no problem with that, but why aren't these types of statistical models and curve analysis being used to deploy drive by patrols? Some of this seems like throwing out the baby with the bath water, does it not? While I agree that it's way too soon for this sort of granularity, is it too soon to have 10 block, increased presence of troubled areas helped by statistical analysis?

AI amplifies human biases. If a few racist cops decide to target a black neighborhood, and "find" a bunch of crimes, that data gets fed into the system. The system then spits out that neighborhood as high risk and assigns extra patrols there.

Those patrols feel the need to justify spending all that time out there, so then they "find" crimes too. And then it just reinforces that data that that is a high risk neighborhood, making the whole thing worse.

The entire system is built on decades of bad data. Step one is cleaning up the data and/or starting over.

Re: Santa Cruz, California bans predictive policing in U.S. first

#170

> Used by police across the United States for almost a decade, predictive policing relies on algorithms to interpret police records, analyzing arrest or parole data to send officers to target chronic offenders, or identifying places where crime may occur. The promise of the software made in a sales pitch or on the website, is a far cry from the reality of what such software delivers. That's fine when it's business, t…

Predictive policing does nothing more than sending policing resources where crime is more likely based on statistical models. It does not racially discriminate and it is not racist. It does not harm anyone. I feel that this is posturing and shooting the messenger. If crime is statistically higher in "black neighbourhoods" the issue will not be solved by pretending it isn't. If the technology does not work then of cou…

Any sort of bias present in the training data will be replicated in the model. If the police are biased in whom they target, that group will naturally show a higher crime rate. Which would easily be picked up in any sort of statistical model. Leading to a biased model.
Post reply on HN