AI vs CPTED at the 2021 ICA virtual conference

Facial recognition technology at a Chinese train station - photo Creative Commons

POST SCRIPT: SINCE THIS BLOG WAS POSTED, THIS KEYNOTE PRESENTATION WAS PUBLISHED BY THE ICA HERE.


by Gregory Saville

A man walks through a public plaza on a pleasant Sunday afternoon and passes by a CCTV. Minutes later he is arrested by police on suspicion of a crime that, in fact, he did not commit. The man is African American. and, unfortunately, facial recognition software on the CCTV is vulnerable to false positives 

A predictive policing algorithm sends police patrols to the same neighborhood for the sixth week in a row to prevent crimes that have not yet occurred. Based on mathematics from earthquake prediction, this algorithm is hardly the best model for predicting human behavior and crime. It has no way to know that residents of this disenfranchised neighborhood are utterly fed up with over-policing, especially when the police don’t actually do anything except show up in their patrols cars. 

I blogged on these stories earlier this year. 

The stories are real and they reflect real events. Unfortunately, according to experts, predictive policing algorithms have serious problems with over-policing minority areas. The Los Angeles Police Department is the latest agency to abandon their PredPol programs (they claim it is due to Covid). Similarly, scientists specializing in evaluation have also criticized facial recognition software. They claim it cannot accurately read facial characteristics of black men! 

These stories reflect the threat of introducing Artificial Intelligence into crime prevention. Thus far, at least with CPTED, things in the AI world are not going well.

The lure of AI in forensic detection and crime prediction

THE 2021 ICA CPTED CONFERENCE

On Nov 3, I will deliver a keynote address to the 2021 International CPTED Association virtual conference, hosted by Helsingborg, Sweden, the Safer Sweden Foundation, and the International CPTED Association. It will be the first ICA conference since the last pre-COVID event a few years ago. The topic of my keynote is Artificial Intelligence, Smart Cities, and CPTED – An existential threat to the ICA. 

Based on my own experience with a tech start-up company a decade ago, and an experiment with some predictive critical infrastructure CPTED software, I came upon some fascinating books on AI. One, in particular, AI 2041 by Kai-Fu Lee and Chen Quifan, describes how AI will infiltrate all aspects of urban life – health, transport, schools, entertainment, crime prevention, and safety.  They tell us there will be no part of the future city without AI. This is especially the case with the Smart City movement in which scientists and planners envision a city embedded with AI. 

Security tech has made inroads into the world of CPTED

THE SORCERER'S APPRENTICE  

What happens when AI systems go wrong? Artificial Intelligence is at the apex of new technologies and the implications for CPTED are significant.

AI is a potential threat of a higher order. It is a case of the Sorcerer’s Apprentice: An independent system that analyses problems and makes decisions using machine learning instructions independent from us. But when things go inevitably wrong, we end up scrambling like mad to stop the damage from unintended consequences (eg:  false arrests and over-policing). 

If you’re interested in this topic in more detail, come to the 2021 ICA CPTED CONFERENCE, which runs from Nov 2 – 4 as a virtual conference. The dynamic conference program has dozens of sessions on crime prevention and CPTED from around the world. My keynote runs on Nov 3 at 9:20 – 9:45 PM Central European Time (1:20 – 1:45 AM Mountain Time). A recording of the conference for registrants will be made available for later watching for those who are asleep in their time zones. POST SCRIPT; THIS PRESENTATION IS HERE.

What's happening to our police? Part 1

Hot of the press: Future Trends in Policing from the COPS Office, PERF, and the Target Corporation. It is a report of a 2012 survey and summary of a one-day session with police leaders on the "Future of Policing."

It reveals what some police executives think might happen in future. Is it a prophesy we really want?

TREND: COMMUNITY POLICING?

The survey reported 94% of respondents said their agency was involved in community policing, 89% in problem-oriented policing (COPS). Good news, right?

I've taught hundreds of police instructors over the past few years. Every time I ask them about COPS few, if any, admit to knowing anything beyond the superficial. Practically none of their agencies are doing anything beyond a small sprinkling of COPS specialists, less than 10% at best.

Last month I asked again, this time whether they knew anything about problem-oriented policing. The class had instructors from the east coast, mid-west, Canada, and the south. Same results: Out of 25 police instructors only 1 knew what POP was and he was from Madison, Wisconsin (the home of the POP Center).

94%?

Do police survey responders inflate whether they are doing COPS when they respond to a national survey on the topic? Saying one thing, doing another?

TREND: SCIENCE TO THE RESCUE

Future Trends had very little discussion of problem-oriented policing. In 45 pages of text it was cited only 3 times.

I did however notice the report was awash in GPS, cybercrime, body cameras, facial recognition software, predictive policing algorithms and intelligence-led policing. My personal favorite was NG 911 - Next-Generation 911.

[NERD ALERT: I love that stuff. Anytime I hear references to Star Trek - The Next Generation, my nerd-o-meter tingles. Beam me up!]

In other words science will come to our rescue? Zip-a-dee-doo-dah, what a wonderful day!

TREND: ONE-IN-FOUR

Survey statement: "In the future agencies will place less emphasis on community policing." 

75% of police agencies in the survey disagreed with that statement. Glass-half-full, right? But 25% offered no opinion or actually agreed that in future cops will do less community policing!

In other words, after 35 years of publications, conferences, training courses, and successes that account for at least some reduced crime, 1-in-4 police survey respondents see less community policing in the years ahead! Sounds more like a glass half-empty!

Considering the Ferguson riots two weeks ago that portends a bleak future.

Next week: Part 2 - The good news

Bad week for predictive policing - PredPol nightmare

Big data to the rescue - Does it work?

Last week was a very bad week for predictive policing, at least one version of it. It comes in the form of an article in the San Francisco Weekly News by investigative journalists Darwin Bond-Graham and Ali Winston. Their exposé, All Tomorrow's Crimes, reported on one company - PredPol - as it markets its computer software to police departments across the county.

It reads like a PR nightmare for PredPol.

It starts innocent enough: "PredPol, short for predictive policing, is riding this wave of techno-mania and capitalizing on the belief... that there's a killer app for everything, including crime-fighting."

I have written on this topic for years. My main worry has been cost and reliability. A year ago I wrote on the predictive model by IBM that Memphis Police call Blue Crush. Two years ago I wrote  The Precog Paradox and three years ago Solving the City With Math.

And now there is a killer app for predicting crime? 

But then the SF Weekly News article takes a darker turn:

"PredPol has required police departments that sign on to refer the company to other law enforcement agencies, and to appear in flashy press conferences, endorsing the software as a crime-reducer — despite the fact that its effectiveness hasn't yet been proven."

Ouch!

I have no idea if any of that is true. But it does show that too many police executives and press outlets are not doing their homework. Where is critical thinking when you need it?

A MEASURED RESPONSE

A year ago a police executive posted on this blog. He is a critical thinker who we need more of in police leadership. This is what he said:

"I sat through a presentation yesterday involving an algorithm-based program that attempts to predict future crime… It could very well be one piece of the problem-solving puzzle. It's something that should not be dismissed. 

It doesn't take the place of a human crime analyst. It doesn't eliminate the need for problem solving. And it doesn't reduce the importance of collaborating with others. It's not cheap, and it could very well tempt agencies to divert funds away from more effective crime reduction strategies…

…only time will tell."

With this news article, it seems it has.

Below is a Ted Talk with a presenter whom the SF Weekly article identifies as a PredPol "lobbyist". It gives you an idea of the PredPol message.

The future is here? And there...

Computer algorithms solving future crime? Photo by WallPapersUS

For ages criminologists settled on the idea that it is not the quantity of officers on patrol that matters most, but rather what they do. Respected police scholar James Q. Wilson (co-founder of the broken windows theory) pointed out it was the style, strategy and behavior of cops that mattered most.

Whatever! Regardless what police scholars thought, the math and computer types spent years figuring how to maximize the quantity and deployment of patrol cars. Squirreled away in dark basement rooms (at least in my imagination), they dedicated themselves to getting the quantity just right.

Disclosure: At one point I haunted those same basement rooms! I was a police planner for a few years and I evaluated PCAM - a computerized program for allocating patrol vehicles into geographic areas.

PCAM - THE LATEST AND GREATEST? 

In the 1980s PCAM was the latest and greatest of its class. It  was supposed to estimate workload peaks, predict hourly call loads, maybe even better deploy cops to cut crime. All very cool stuff. It didn't work of course. PCAM just couldn't handle high priority calls. Still, it was pretty cool.

Unfortunately, no police allocation model received any level of wide acceptance throughout the 80s and 90s. That didn't change the fact that knowing where crime will happen is no small feat and so those basements remained very busy places. In fact, research continues today.

I've written all this before in blogs on predictive policing and the Precog Paradox.



THE BRITS GET IN THE ACT

Now for the latest! It looks like 007 is asking about the most recent variation on that old theme: Predpol. And compared to the others, Predpol is even cooler - in a shaken, not stirred, kind of way.

Predpol has crossed the Atlantic to the Kent Police. The Brits, it seems, are grasping at the same quantity straws that we are. Let's hope they don't obsess on where crime happens and ignore programs and funding on why they happen in the first place!

Sci-fi policing and the Holy Grail

"When predictive experts fail they are just replaced by a new group who say they can do better." John Ralston Saul

The prediction game is making the rounds again. If we are to believe the New York Times and the Huffington Post, there is this tiny corner of criminology where computer scientists and math types squirrel away like mad scientists decrypting secret code.

The predictive experts are the latest media darlings. No facetiousness intended. If prediction delivers on what it promises and becomes an early warning sign to improve community-based, proactive problem solving you can count me in.

IBM and Memphis Police

The latest iteration, this one from IBM, is on a crusade and they have some new friends in police departments like Memphis.  Memphis points to their new predictive policing program called "Blue CRUSH" to account for a 26% crime drop in the past five years.

Sci-fi policing in real-time! Kewl. The mad scientists are positively tingling. IBM's background report explains it all.

Wait! Not so fast.

From 2002 to 2011 Tampa had a 72% felony crime plunge. Police say Tampa did it with proactive problem solving, analysis from their crime mapping unit, and a compstat performance review to hold area managers accountable. Others point to demographics. Tampa had a 40% increase in 50-64 yr olds with the financial resources to trickle some positive mojo into the local economy. That in turn mitigated, or displaced, younger crime-prone age groups.

Either way, predictive policing had nothing to do with it.

Predictive policing and the promise of sci-fi

Still, I have a soft spot for the sci-fi promise. My blogs show it; Solving the city with math and Predictive policing and the PreCog paradox are the latest examples.

Disclaimer: In 1988 I co-published a predictive spatial analysis on probable locations of professional auto thieves. In the 1990s we expanded that into a tipping point theory predicting how neighborhoods tip into crime. Sci-fi policing groupie? Guilty as charged.

All great fun. All beside the point.

Police resources nowadays, razor thin and bloated on salaries, can scarcely afford expensive math experiments. Tampa did fine without it. Would they do better with it? Maybe. But if demographics are the primary cause of crime declines then we're fooling ourselves with fancy math and ignoring the root social causes that trigger it in the first place.

Anyway I'll continue reading about it. After all, who can resist such appealing titles like Poisson-based regression analysis of aggregate crime?

Not me.