A predictive model isn’t always trying to predict the future. It can also be used to predict the present—to find hidden connections in (perhaps maliciously) incomplete data.

In this case, that’s to say: if an unnamed gang member committed a crime, but all the known members of said gang have alibis... maybe you should be looking for an unknown member of said gang. And who’s likely to be secretly in a gang? Well, someone whose friends are all in said gang would be a good first guess.

OMG.

What about outreach workers? How many of them are going to have connections to junkies, homeless people and gang members?

Sincere good Christians often believe in trying to help the unfortunate. Some good Christians spend all their time surrounded by clean living church goers. Others spend a whole lot of time surrounded by folks who are from the wrong side of town and the wrong side of the law.

The point of these analyses isn’t to say “arrest this man he’s a criminal” but rather “closely observe this man and ascertain whether he’s a gangbanger or an outreach worker.” Of course it doesn’t even need to get that far since it’s trivial to tell the two apart from the network analysis alone (the outreach worker will have a far higher ratio of legitimate-world contacts to criminal-world contacts than the gangbanger).

And you don't think law enforcement agencies will quickly make the leap from "closely observe" to "harass and arrest for any tiny transgression in an effort to coerce a confession"?

No, because they're overworked as is.

Exactly. Instead of doing their job manually, or double- checking system output, they begin to take what the software package says as truth. Whether it’s just human nature, or overwork, or whatever, it’ll happen.

I don't, no, why would they?

---

Downvoters, care to explain?

Because Law Enforcement generally behaves like people, and (like people), often take cognitive short-cuts when making certain assessments. In this case, the short-cut would be to assume that the individual has to be a criminal, and act as such.

Now, there are step that you can take to reduce these assumptions, but those are steps outside of the program, and would have to be introduced in tandem with the introduction of the program. Unfortunately, policy-makers often assume that any single idea that is funded (i.e. buying a predictive model from Palantir) is a comprehensive solution to the problem (often encouraged by the sales reps peddling whatever the 'solution' is) and fail to recognize that other programs will have to be funded alongside in order to deploy the program effectively.

I am law enforcement - the opposite is true. LEOs are looking for any opportunity to close off their cases with as little effort as possible.

If the software says Bob probably did it, Johnny Law will assume Bob did it and stop looking elsewhere, precisely because Johnny Law wants to close the case. Throw an overzealous DA in the mix, and it sucks to be Bob.

That's just not how it works. Eventually you have to present actual evidence in an actual court and it won't be "this software you never heard of said so".

No downvote interest, but US history is overflowing with consistent, wrongful police harassment based on socioeconomic and ethnic lines.

There is 0 reason to expect them to not abuse this

If you study social psychology, history, etc, life generally does not agree with your conclusion.

“Have the police scrutinize the outliers” is a fantastic policy if you are fine building a society that can’t support building a community like HN.

Seems like an unfair assumption. "Legitimate-world contacts" could be criminals too.

Maybe you did not get the memo, but there are, for example, pedophile priests in the world.

Methinks fine upstanding citizens are not as readily distinguished from bad guys as you seem to imagine.

The word ’predict’ gets used a lot and is sufficiently charged that it causes much Minority Report-style angst and gnashing of teeth, but really a better term would be ’uncover’. The whole investigative process is about uncovering concealed connections between people and events, so this is just a new technique. It isn’t inherently Orwellian either: there’s still a wide chasm of human investigators, prosecutors, judges, and juries between any kind of ’positive’ and actual consequences.