Chicago PD's Big Data: using pseudoscience to justify racial profiling

I was just saying that (a) systematically discriminating against people makes it more likely that they will run afoul of the criminal system so (b) if you have algorithms that purport to show the risk, and systematic discrimination based on the algorithms, the discrimination will reinforce the algorithm.

I guess you could call it a self-fulfilling prophecy, except for all the agencies, businesses, and so on doing the work of fulfilling it.


We had a similar system over here on the other side of the pond seventy years ago. Some pseudoscientific data pointed out a specific group of people that were more likely to be criminals. They had to wear little stars on their clothes.

With hindsight it was not a good idea.


How do you know that it’s a fact that it’s scientifically valid approach if the algorithm is opaque and not subject to confirmation or verification? Security by obscurity is neither security nor science.


Yeah, Zoroastrianism had a couple of weird outcomes all right; diversity gone foldover pongy.

Is this a crime, you dystopian, Orwellian fucks? Did you predict it?


In person? Yes, actually. Flip off a cop? They charge you with “Disorderly Conduct”.

Sounds too insane to be true, right? …right?

1 Like

Yeah, stop down a little and pull focus on something, say either the corduroy or the fur…whatever that is. Is it disrespectul to the greenscreeners’ craft to not key on it, then shoot the studio gamut with daylighting ablaze? Toggle the bit to put a magical land behind you already. ITV studios, say.

1 Like

Freddy Mercury, maybe, but debatable. Kaikhosru Sorabji, definitely. Zubin Mehta, not so much. Lots of diversity there, though…

1 Like

^800x1000px, if anyone wants it


Your number has come up…

Here’s an example of how this could be (and likely is) pseudoscience:

  • we know that whites use marijuana at a higher rate than blacks
  • we know that those have gone to college use marijuana at higher rates than those who have not
  • we also know that blacks are much more likely to be criminalized as a result of marijuana use than whites
  • what we don’t know is how CPD is determining risk. If they are using usage rates and who is actually likely to break the law, we would expect a lot of white male college kids and their parents to be getting these courtesy calls. If they are using arrest and conviction rates, then we are going to get a lot of uneducated black youth getting these little visits. I suspect that if the CPD really are visiting these white college kids to warn them of their imminent criminality and future criminalization, then their program is going to have all of its public support bled from it pretty quickly.

I heard Miles Wernick has a version of this for Wall Street criminals.

I mean … he does, right?

Hello? Hello?


He compares it with epidemiological approaches, stating that people whose social networks have violence within them are also likely to commit violence.

There are so many problems in that one sentence that I hardly know where to start. I’ll try to keep it short, and just say that I’ve spent thirty years developing software, and the last fifteen years of that time I’ve spent helping scientists, many of whom are epidemiologists. I am not an epidemiologist, but I’ve worked extremely closely with them, and I’ve learned that a great epidemiologist will be the first one to tell you that, like most areas of scientific study, epidemiology is so easy to do poorly that you have to be on constant watch for what amounts to dangerous bullshit. Having spent most of my life working with people that have advanced degrees, and working at major research universities, I can assure you that people with advanced degrees and from major research universities have just as many bone-headed ideas as everyone else. In particular, I noticed very early on in my career that a startling number of people with advanced degrees develop a very myopic and narrow view of the world, where they might get an idea they believe is great, and fail to see obvious privacy, ethical, logistical, statistical, or many other issues, because those issues are outside of their field of study. Ideas like this. We now have a militarized police force that has essentially unlimited funding from Homeland Security and “drugs”, and they will take cover under any idea, no matter how ludicrous, to justify their behavior.


The claim that it’s pseudoscience is pretty strong. Lots of arguments were put forward to support it.

Can you put forward a single argument that it is actually science? Because so far the only one you’ve put forward are religious arguments, not scientific ones. (The qualifications of it’s “champion” and the popularity of the approach).

The extent of your argument basically seems to boil down to “I don’t understand the difference between science and mysticism, so this must be science!”.


The lack of disclosure really is understandable. The Pre-Cogs need their privacy, after all.


It may be bad science, but the concept makes excellent science fiction. “Time Considered as a Helix of Semi-Precious Stones” by Samuel R. Delany won a Hugo and a Nebula in 1970. It begins with a petty criminal being visited by a special investigator because her database says he will commit a crime, even though she doesn’t know which or precisely when. Neither does he. Good read.


Not a criminal.

Followed by police, who told me I was probably a criminal.

May as well be a criminal, they say I am.

Conviction & incarceration rate secured. Budget process sealed.


The only fact had about the scientific process is that the public is told it is one.

Science Wept.

1 Like

Don’t worry, something tells me employers, insurers, and lenders already operate using the same datasets Cory is talking about.

1 Like

Because people act so rationally on such a regular basis.

Also, in the gimme-a-break category, The Verge’s FOIA request about the program:

...was denied because sharing that information could "endanger the life or physical safety of law enforcement personnel or any other person," according to a letter from the CPD’s Office of Legal Affairs.
See, because then the folks in the ghetto (:notes: in the ghetto :notes:) would build their own algorithm so they'd know the next house where the cops would show up and they'd be waiting!