I've written about how we did it in Oak Park, IL:

* https://news.ycombinator.com/item?id=40227280

* https://news.ycombinator.com/item?id=41927777

* https://news.ycombinator.com/item?id=45506690

Cards on the table that I was not a full-throated supporter of cancelling our Flock contract, for complicated reasons, but past that I'll take a fair bit of credit for the harm-reduction work we did, which ultimately created the procedural tracks we used to kill the contract.

Short answer for how we did it: message board nerding.

You're interested in getting the cameras taken down in a Wyoming muni. One advantage we had in Oak Park that you might not in WY is that our cost function priced bogus stops of Black drivers very high. So, if I was strategizing killing cameras in a major metro suburb, my strategy would be:

(1) Create procedural rails to collect your own transparency data on stops.

(2) Do the analysis to trace "real" stops to crimes meaningful to your muni (for us: enforcing failure-to-appear warrants for neighboring suburbs was not high-value work for OPPD, so many of the "legit" stops had negative value).

(3) You'll be left with some subset of real crimes cameras were involved with, and in only a subset of those will the cameras have been meaningful.

One thing that complicates Flock deployments in Illinois is that they depend on the ISP LEADS database as their hotlist of stolen vehicles, and LEADS is not maintained well enough to use as a real-time information source (or even a week-by-week granular source), so we had a bunch of bogus stops.

A super important thing I think everybody should know about Flock cameras:

You do not in fact need to be enrolled in Flock's sharing system to get data from neighboring muni cameras. In fact, I think Flock even has a product you can buy that just gives you access to sharing data without even owning cameras.

Since "we need to share our data to get access to other muni's data" is the only reason to have sharing enabled on these things, it should be pretty easy, as a political lift, to turn sharing off.

▲

aerostable_slug 7 hours ago | parent [-]

Why would a stop for an outstanding warrant have negative value for your community? If word got out that OPPD will come down hard on anyone with warrants, perhaps people with warrants would stay away from your community. I'm not sure I see the downside; deterrence is a good thing.

Analogy: criminals know Target stores have a policy to prosecute all shoplifters, so when there was still a shoplifting subreddit that fact would be regularly trotted out and criminals were warned by their peers (the best kind of testimonial) to stay away. I would love it if my neighborhood had that reputation.

▲

tptacek 6 hours ago | parent [-]

A failure to appear warrant is generally someone not showing up to court to pay a traffic ticket. It's essentially municipal debt collection work. I'm not saying it's bad to catch people with outstanding warrants; i'm saying that OPPD curbing a car and making an arrest has a simple logistical cost, and that cost swamps the minuscule value of helping a neighboring suburb collect ticket revenue.

Our police have real work to do. If we had a special magic beepy device in all the police cruisers that lit up when someone with an outstanding warrant drove past, we would not prioritize that enforcement work to the exclusion of the real work. But since OPPD doesn't know that they're going to end up burning 5 hours on a failure-to-appear warrant when they curb a car on a Flock alert, that's what Flock essentially had us doing.

I honestly think this argument is probably pretty portable to a lot of different municipalities. It's not a function of anything Flock itself deliberately does, but rather a simple function of pretextual or preemptive stops on cars: you are probably going to end up making a whole bunch of failure-to-appear arrests. And I think in pretty much every community where killing camera contracts is on the table, failure-to-appear enforcement will be perceived as net-negative, a distraction from preventing serious crime.

The thing I like about this argument is that it's insensitive to people's priors about law enforcement. Whether or not you like your PD (I very much like OPPD), this argument should have weight!

The key observation here, again, is that any arrest has a very high fixed cost.

▲

loeg 4 hours ago | parent | next [-]

I don't know, I think paying your traffic tickets is about the least you can do downstream of very occasionally being caught for habitual dangerous driving behavior. Breaking the "municipal debt collection" breaks the deterrent effect of traffic tickets.

I agree that in the abstract maybe there are better things some cops could be doing, but it seems like a vaguely reasonable use of some traffic enforcement resources. It's not like this taking away from murder investigations.

There's a prisoner's dilemma defecting thing going on here, right? You'd want neighboring municipalities to enforce warrants out of Oak Park.

▲

tptacek 4 hours ago | parent | next [-]

It literally does take away from violent crime investigation! Remember, I'm not making a moral argument about the legitimacy of traffic fines. In fact, that's one of our big issues in Oak Park. I'm saying that police departments make prioritization decisions, and Flock cameras structurally undo those decisions by throwing alerts on cars (which would not otherwise have been curbed) that produce warrant arrests.

The key thing to understand is that an arrest eats half an OPPD officer's work day, so if OPPD is arresting someone, you want the juice to be worth the squeeze.

▲

loeg 3 hours ago | parent [-]

Your traffic cops would otherwise be participating in murder investigations? My understanding is that these are different specializations and they don't overlap.

No real objection to "the data source is bad," but I think the solution there is improve the data source rather than willful blindness.

	▲	tptacek 2 hours ago \| parent [-]
		I didn't say "murder investigations". We have a very small number of detectives, who do not conduct traffic stops, but the overwhelming majority of our force (and of all the police work done here) is patrol, all of which do conduct stops. We don't control this data! It's good to want things, but whether or not you think it's good that LEADS isn't good enough for real-time enforcement, it is not.

▲

queenkjuul 4 hours ago | parent | prev [-]

I think lots of upstanding citizens have forgotten to pay a ticket at some point (as in, I've done it, as have some other people i know, and none of us are criminals or even dangerous drivers. My ticket was for a one-day expired license plate sticker, btw). The cop doesn't even know from the license plate that the person driving is the owner of the car, and so they don't even know that the driver actually has a warrant until after they're stopped.

▲

aerostable_slug 5 hours ago | parent | prev [-]

That's interesting, and makes sense.

Mr. Flock Person, how about a feature request to alert only on 'interesting' warrants? I wonder if that's even possible — it might be a binary flag on the plate data. Hm. If so, that would be a serious bummer and something perhaps the legislature should look into remedying (such a change would require funding after all). A will-extradite flag seems like it would be useful.

▲

tptacek 4 hours ago | parent [-]

I know there's Flock staff commenting on this thread (which is great, regardless of how you feel about Flock) but just for context: Flock alerts in Chicagoland municipalities are driven by LEADS and the lookups are on license plates, not people, so they can't actually filter out failure-to-appear stops.

	▲	aerostable_slug 2 hours ago \| parent [-]
		Right, that's why I was thinking the legislature would have to get involved. An extension to available license plate data would require approval and funding. I think law enforcement everywhere in the US would like a way to make ALPR more useful for targeting higher risk offenders and leave the revenue generation warrants to times when the subject is being arrested for a different offense. Such an extension would maximize the utility of the investment that's already been made in ALPR hardware, software, and services.