LAPD Watchdog Says Department's Data-Based Policing Is Producing Nothing But Wasted Time And Rights Violations
from the only-looks-like-you're-working-smarter dept
The Los Angeles Police Department has just received some bad news from its oversight. It's probably good news for the policed -- many of whom are being disproportionately targeted thanks to biased input data -- but the LAPD can't be pleased that its reliance on expensive, mostly-automated tools hasn't produced worthwhile results.
The department relies on a handful of tech tools to aid in its policing, but it doesn't appear to be helping. It has CompStat -- a holdover from the early 2000's when Bill Bratton still ran the department. To that framework, it has added LASER -- a nifty acronym that stands for "Los Angeles' Strategic Extraction and Restoration." The program with the reverse engineered nickname actually relies on input from human analysts to determine where officers should be deployed. But this reliance on data-driven policing isn't making the city any safer, despite LASER's focus on violent crime.
Here's what the LAPD's human analysts put together for the department's patrol officers.
In perhaps the most contentious strategy, each of the department’s 21 geographic areas used data to compile lists or “bulletins” of people calculated to be among the top 12 “chronic offenders.”
The program assigns people points based on prior criminal histories, such as arrest records, gang affiliation, probation and parole status and recent police contacts.
This strategy received some public blow-back, resulting in the department abandoning it last August. Nothing of value was lost.
[Inspector General Mark] Smith examined data collected prior to the suspension.
He found 44 percent of chronic offenders had either zero or one arrest for violent crimes. About half had no arrest for gun-related crimes.
So much for curbing violent crime. All it did was create a loop where cops targeted nonviolent offenders, resulting in another arrest/detention that added more points to the person's LASER record, resulting in even more targeting and, inevitably, more interactions with police officers. It's a feedback loop no one can escape.
To make things worse, officers had the power to place people into this damaging loop by "nominating" them for targeting with LASER. The point-based system that was supposed to limit this targeting to just the worst of worst street criminals could be bypassed. Nominated citizens would find themselves rising up the ranks on the LASER lists, racking up points simply by officers performing stops based on faulty inputs.
And while the tech is supposedly improving, the quality of policing isn't. CompStat has had nearly a 20-year run in LA, but its results are negligible. Predictive policing -- which has its own bias issues -- isn't doing any better.
Like the other program, Smith found discrepancies with the data collection and could not draw conclusions to “meaningfully evaluate” the program’s overall effectiveness to reduce crime, the report said.
Unfortunately, the report recommends the LAPD stay the course. The LAPD is supposed to spend more time "reviewing" the data that isn't producing results and tailor its outputs with an eye on Constitutional rights. As it stands now, the LAPD is allowing databases to conjure up reasonable suspicion for stops. That can't keep happening. But the way forward can't be more of the same, only at a slightly slower pace.
Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.
Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.
While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.
–The Techdirt Team
Filed Under: civil rights, compustat, data-based policing, lapd, laser, los angeles, mark smith
Reader Comments
Subscribe: RSS
View by: Time | Thread
The lack of understanding of the criminal mind, or any mind
Does this tell us that humans don't do any better at 'pre-crime' analysis than algorithms will? Humans write algorithms, don't they?
Predicting crime might be possible, someday, but taking action on those 'predictions' should take a lot longer. We might be at the beginning of understanding the criminal mind, but predicting what their next action might be is something else. Then there is the difference between the long term, repeat offender, and the new, not yet known offender in the making. Predict them!
I have yet, after years of interaction, found anyone who could tell me what I am thinking or what I am about to do. Part of that is I don't always know what I am about to do (I like surprises) and part is that no one is absolutely knowable. Criminal or not.
[ link to this | view in thread ]
Can the citizens of LA do the same to the cops?
If the people of LA start their own program of most distrusted police and nominate officers who show violent or anti-ethical behavior, would that lead to those cops being prosecuted instead of protected behind the thin blue line?
[ link to this | view in thread ]
Re: Can the citizens of LA do the same to the cops?
Probably not, but the embarrassment coming from the disclosures of bad behavior by law enforcement officers that are eminent (that is if they are not destroyed first, but that is an entirely different legal matter that if it becomes truet I predict will go nowhere after years of huffing and puffing) might be enough to do some damage to some officers for past behavior. Of course they should have had that damage when those events occurred.
Part of the problem is that the cops are on the prosecutors/judges side, and they don't like to have penalties thrown against them. Then there is the ability of the officers to amend their statements, over and over, until they get it 'right' because they don't follow the same rules as us peons. Then there is the statute of limitations, whereby some, if not many offences that were not prosecuted, or not prosecuted properly will go without any detrimental effect to the perpetrators who have had significant vacation time (erm paid administrative leave) and their legal bill paid by their union.
[ link to this | view in thread ]
Re: Can the citizens of LA do the same to the cops?
Gotta get the unions and AG to stop claiming the recent transparency law is not retroactive first.
[ link to this | view in thread ]
Statistically Speaking
"Nothing But Wasted Time And Rights Violations"
<duh>Let's all feign surprise at what poor statistical results you can get when you don't bother to consult any professional statisticians.</duh>
[ link to this | view in thread ]
Re: Statistically Speaking
yeah, LAPD buffoons are clearly the problem here -- not the concept of statistical analysis of crime patterns & individual criminal history.
[ link to this | view in thread ]
Re: The lack of understanding of the criminal mind, or any mind
How much crime maps onto people in areas in economic desperation, with poor education, chances of a job and housing? Fixing that is a political problem, and not a policing problem.
[ link to this | view in thread ]
Re: Re: Can the citizens of LA do the same to the cops?
Let's not even go as far as something that needs to be prosecuted. How about just acting like a dick?
Yesterday, I watched a sheriff flip his lights on for a couple of seconds for one tiny silly reason. There was an 80 year old guy driving barely under the speed limit in the left lane, and the cop couldn't be arsed to go around him like a normal person. The road wasn't even crowded, but this cop tailgated the guy for half a mile, I guess waiting for the old man to respect his authoritah. Cop flipped on his lights, old man swerved into the right lane, and cop went on his merry way at 5 or 10 over the speed limit (lights off). That's just being a jerk because he could.
[ link to this | view in thread ]
Policing is biased, this is a well established fact backed with tons of data. In order to save face, because they are that vain, the politicians have been wracking their brains for a way to blame others for their bigotry and now they have their scape goat ... software. Do not blame me !!! The computer did it!!!
[ link to this | view in thread ]
Re: Re: The lack of understanding of the criminal mind, or any m
The police think the crime committed by a poor person is a bigger threat to society (the rich) than the ever present white collar crime.
If they were to conduct stop 'n frisk on Wallstreet they would nab a lot of druggies, but they are not interested.
[ link to this | view in thread ]
Re: Can the citizens of LA do the same to the cops?
"If the people of LA start their own program of most distrusted police and nominate officers"
They will find themselves on the receiving end of police harassment (they make shit up).
[ link to this | view in thread ]
Re: Re: Statistically Speaking
Who is directing that the stat software be used?
[ link to this | view in thread ]
Re: The lack of understanding of the criminal mind, or any mind
Garbage in, garbage out.
Computers aren't magic. You can't plug in biased data and get an unbiased result. If the problem is that laws are not being enforced fairly, you can't change that by feeding existing enforcement data through a mathematical formula.
[ link to this | view in thread ]
Re: Re: Re: Can the citizens of LA do the same to the cops?
If it was a limited access highway, the cop was being nicer than he needed to, since being in the left lane and not passing is a violation in most states. If a normal surface street then that was a dick move.
[ link to this | view in thread ]
Re: The lack of understanding of the criminal mind, or any mind
"Predicting crime might be possible, someday, but taking action on those 'predictions' should take a lot longer."
Predicting crime is impossible.
Taking action(s) on the predictions would be silly.
I can predict that when a politician opens their mouth, what comes out is a lie. This is backed by evidence and statistical analysis, but this is not infallible. We already arrest/harass too many people for no reason, is this simply the excuse for doing so? Because one can certainly make it say whatever they want it to.
Figures do not lie, liars figure.
[ link to this | view in thread ]
Re: Re: Re: Re: Can the citizens of LA do the same to the cops?
The guy driving a bit too slow is prob the result of a prior cop dick move issuing a ticket for 3 miles over
[ link to this | view in thread ]
Can't escape?
Correct me if I'm wrong, but couldn't one stop committing crimes to avoid it? Is the suggestion that the cops are rearresting these people for no valid reason or on a flimsy pretext (like something from Three Felonies a Day)?
[ link to this | view in thread ]
Re: Can't escape?
Yes.
We're talking about police performing stops based on "reasonable suspicion". Techdirt has written quite a lot about that subject; here's a quick search to get you started. "Flimsy pretext" is about right.
[ link to this | view in thread ]
I told them over and over
that LASER wouldn't work. I told them to use SCMODS. But, did they listen to me? No!
[ link to this | view in thread ]
Re: Re: Re: Re: Can the citizens of LA do the same to the cops?
That actually...doesn't make sense for many urban limited access highways, as they aren't operating at a Level of Service that would permit the left lane to be a passing-only lane (never mind what you do when you have 5 lanes of traffic in each direction).
In other words: if the facts that justify having a law present on the books don't hold in a circumstance, trying to enforce that law is daft at best and dickish at worst.
[ link to this | view in thread ]
Attribution requested
You state that "And while the tech is supposedly improving, the quality of policing isn't. CompStat has had nearly a 20-year run in LA, but its results are negligible. Predictive policing -- which has its own bias issues -- isn't doing any better. " Nowhere in the article is any justification for that statement. You have no more statistics to justify your position than the cops have for theirs.
[ link to this | view in thread ]
Re: Re: Re: Re: Re: Can the citizens of LA do the same to the co
I guess such laws are not applicable during rush several hours where it is bumper to bumper fifteen miles an hour in all lanes
[ link to this | view in thread ]
Re: Re: Can't escape?
Welcome my son, welcome to the machine.
[ link to this | view in thread ]
Re: Re: Statistically Speaking
My predictive software projected a 91.4% probability that someone would use your IP address to make that remark.
[ link to this | view in thread ]
Re: Attribution requested
Article linked in this article from the LA Times:
https://www.latimes.com/local/lanow/la-me-ln-lapd-data-20190312-story.html
Inspector General Mark Smith’s 52-page review found officers used inconsistent criteria to identify people with criminal histories who are most likely to commit violent crimes. Department policies should “unequivocally state” that officers must have legal justification in all situations before confronting someone, Smith warned.
Further down, the article TechDirt linked to talks about CompStat. And then has this portion:
Like the other program, Smith found discrepancies with the data collection and could not draw conclusions to “meaningfully evaluate” the program’s overall effectiveness to reduce crime, the report said.
And you'll note that that exact quote is in the TechDirt article. The supposition that results are negligible is based on the statement from the IG.
In addition, when first talking about CompStat, an earlier TechDirt article is linked:
https://www.techdirt.com/articles/20190124/15171141462/criminal-justice-system-is-relying-te ch-to-do-job-thats-just-going-to-make-everything-worse.shtml
This article, among other links, has a link to the following:
https://www.technologyreview.com/s/612775/algorithms-criminal-justice-ai/
This is the Massachutts Institute of Technology, btw. That article is worth a read - it informs a lot of what's put into this article.
Remember, AC - follow the links. They aren't linked for giggles.
[ link to this | view in thread ]
Re: Re: Re: Re: Can the citizens of LA do the same to the cops?
Nope, not a limited access highway.
California doesn't have that kind of lane control either.
Just a cop saying "Excuse you!" to someone driving in a more legal manner than he was.
[ link to this | view in thread ]
Re: Attribution requested
When you see blue text on the Internet, that's usually something called a hyperlink, or just link for short. If you click on a link, it will often take you to another page with different content.
[ link to this | view in thread ]
counter-intellugence tactics, targeting civilians
re: “LASER....The program with the reverse engineered nickname”
in linguistics, this is called auto-antonymal language.
It is one of the most common features of counter-intelligence psychological operations, aka organized gang stalking.
Or, what the Association of Threat Assessment Professionals(ATAP) calls a "colliding parallel investigation,”aka organized gang stalking.
https://c.ymcdn.com/sites/www.atapworldwide.org/resource/resmgr/2018-ATAP-Preliminary-TMC- Pr.pdf
[ link to this | view in thread ]
Re: Can't escape?
What an arrogant, ignorant, bizarre presumption. You miss the point entirely.
Yes, newsflash the police target, and prey upin, and stalk the poor, and the lower classes.
Its a numbers game, policing for profit. The weak/vulnerable/overly policed get pulled over and stalked by these paid armies of elite criminals who violate every single word of the constitution, every day, with these Fusion Centers and associated parallel construction schemes.
So, your comment is either willfully ignorant of class oppression, or you are on yet another donut run in your state funded cruiser, or police union purchased SUV.
[ link to this | view in thread ]
Re: Re: Re: The lack of understanding of the criminal mind, or a
They would nab the druggies, but they wouldn't get any convictions because wallstreet druggies have the money to get criminal cases tainted by fourth amendment violations tossed and they have their lawyer telling them what to say to the cops (nothing) and they have the money to post bond and walk away instead of rotting in jail while their life dries up and blows away.
[ link to this | view in thread ]
Re: The lack of understanding of the criminal mind, or any mind
"Predicting crime might be possible, someday, but taking action on those 'predictions' should take a lot longer. We might be at the beginning of understanding the criminal mind, but predicting what their next action might be is something else."
Well, there IS one action we can predict taking place in the criminal mind...
...the very second such a database sees preferred use by any law enforcement agency some hacker or less upright officer of the law will be handing over the algorithm and database to an organization which uses it to find out where the police will not be present at any given moment - and then sells that information to local gangs.
The assumption that a predictive database will help rather than hinder assumes - with guaranteed failure - that said database will only be in the hands of law enforcement. Once it leaks it's suddenly a force multiplier for the criminals instead.
And that's even before we go into the details on whether such a database could ever be made to work. Political stupidity at its finest.
[ link to this | view in thread ]