Bring on the cameras. I give up. I’m ready. Recently in Chicago, a man strong-armed a woman sitting at a bus stop into an alley and sexually assaulted and robbed her. This happened at two in the afternoon on a clear day in broad daylight at a very busy intersection. I know it’s a busy intersection because I live two blocks away.
So I want this guy caught. I want people to know that crimes committed in public are going to lead to arrests. I want a camera grid to do it. But I want it implemented smartly.
What We Have
What we have right now is a low quality system with no protections.
As we saw in the recent Boston Marathon bombings, in a way a camera grid system is already in place. It’s a de facto system that is less effective and has no protections on abuse. Police, relying on the good will and availability of nearby businesses, simply ask for the footage they need. This footage can be of varying quality reducing effectiveness, it can be of varying formats expanding the time it takes for police to process once it’s in hand, and it is held for varying amounts of time meaning it may not be there when needed.
A rogue official (or agency) willing to abuse her (or its) power could ask businesses for footage for whatever purpose, making up whatever excuses necessary. Some places could ask for warrants or other proof of need, but there are no guarantees. On top of this, there are no official records of who asked for footage, when, and why. Why leave this up to chance?
Our ad-hoc system also poorly handles crimes committed in residential or lower income areas that don’t have businesses nearby with cameras for security. There simply won’t be coverage by cameras in some places within a city.
What I Want
Camera technology is improving; smaller cameras capture better quality images than ever. Cameras are already ubiquitous, and fighting against that ubiquity now would be futile. Privacy concerns are valid, but they don’t change our new reality. It’s better to responsibly harness the power of cameras for good and to get out in front of these tough privacy questions.
Firstly, I want the camera grid system to be useful. It should capture a clear, usable picture of an offender’s face and it should be able to reliably track a possible offender as they move throughout the city. Any outlay for a system that can’t at least do these two things would be a waste.
A second requirement would be license plate recognition. Yes, this almost sounds even scarier than facial recognition. In a car-dominant city like Chicago, it is necessary to track cars in order to track suspects. The license plate tracking would be subject the same restrictions as the regular cameras, with an additional list of “hot” plates to watch for. This plate list would require case numbers just as the regular cameras do, its access would be audited, and those audit logs would be available to the public for oversight. Currently, some municipalities have such devices, but the oversight is poor to say the least. While we’re passing a law for a general camera system, let’s get this more specific technology under control as well.
A third requirement would be that the camera grid covers the whole city. Some areas, like low-density residential or industrial areas, could certainly have fewer cameras per block, but in high density areas we should have cameras blanketing as much as possible.
Economic Impact and Costs
Of course the program will cost money. I can’t go into detail on how much, but storage and cameras are getting cheaper all the time.
More interesting to look at are the economic benefits of creating a safer city. Estimating the costs of violent crime is difficult to do, but that doesn’t mean it’s impossible. At least one study puts the cost for the City of Chicago at $1.1 billion dollars per year. The increase in the municipal budget for a 25% decrease in violent crime would be $49 million dollars per year. If a camera grid can aid in that kind of crime decrease, the return on investment would take years, not decades, certainly. The $49 million is strictly the government budgets, local businesses and residents would reap their own benefits.
Why It Might Not Be So Bad (Privacy Concerns)
There are two broad categories of privacy concerns – those of the rogue agent and those of the police state. How can we alleviate these concerns?
It should not be run by humans. Very little human interaction should be required for day-to-day usage, and access to particular video feeds should be restricted. Only a law enforcement official pursuing information about a reported crime could access the system, and the access would require a related case number. Information on who accesses the system and when should be available to the public. I’m split on if this should be generally available or by FOIA request, but it’s a mandatory requirement.
This is a critically important point, and worth considering further. The primary curtailment on abuse is this open view into use of the system. A rogue agent entering bogus case numbers to view feeds should be relatively easy to detect. The more difficult issue is when the state or system itself attempts to hide behind the “ongoing case” exception to a FOIA request to monitor a non-criminal person, agency, group, or gathering. But the cops are already monitoring non-violent groups discretely (sometimes very poorly). They’re going to monitor non-criminal groups no matter what, let’s at least make them work for it.
I should add that this idea of an audited surveillance system was planted in my head via the book The Transparent Society by David Brin. Strangely, this book is the only place I’ve seen the idea mentioned. Politicians and privacy activists seem to only concern themselves with the all-or-nothing extremes, never concerning themselves with a practical, viable suggestion to handle the issue while minimizes the fears and problems it presents. Media stories on these items always leave out the third way options as well.
With this system in place, when a man assaults and robs a woman in broad daylight we can now get an accurate picture of his face, we can find out where he came from on transit or by car, and where he went afterwards. If he did drive a car before or after, we have his license plate information as well. All of this knowledge is available quickly to police in a system designed for such pursuit, with audited access so that we, the public, can confirm its being responsibly used. Whew.
A Smaller Example
Property crime in Chicago is commonplace. For items worth less than one hundred dollars, there is an automated report system; you don’t even talk to a cop if you’re a victim. I’m sure this is half due to the volume of the crimes, but also half due to the nearly impossible task of tracking down a single thief in a city with a population of millions. These sorts of crimes would become easier to solve, assuming the police had the manpower to task people to tracking the offenders down.
I’ve purposefully avoided technical details of the plan, but there is one thing I believe is worth talking about. I’m not a security expert, but working around computers as much as I do I know that it’s nearly impossible to secure a system against someone who can physically access the system. As such, no Chicago PD member should be able to touch the systems that run the grid and store the data. Since it’s critically important data, it should be secured somehow, but by an outside organization. I’ll throw the Illinois State Police out there as an idea, but this is a detail to work out later.
An Eye on The Future
My understanding is that the technology isn’t quite there yet – but integrating the system with automated behavioral detection algorithms should be easy, and the system should be designed with this in mind. If the system can detect possibly criminal behavior or noises and alert police before a crime has occurred, this would be superior to a system that only assists with easier capture of criminals after the fact.
Facial recognition is also improving, and I could see the possibility of adding pictures of persons with outstanding warrants to an automated watch list – again something carefully and publicly audited to avoid abuse. Alerts would be automatically sent to officers in the area when a wanted person is detected nearby.
I know such things sound scary, but the technology is coming. Denying it won’t help, but thinking about the pitfalls and working to head them off could.
Let’s Solve Crimes
I’ll finish with a quote from a story of the assault and robbery that happened down the street from me:
Even though the crime happened in broad daylight in a very busy intersection, no one saw anything. A diner is less than 10 feet away from the bus stop and the people inside the diner said they did not see anything, although they did say at 2 o’clock in the afternoon, the diner is usually empty, as it is their least busiest time. There are also no surveillance cameras outside.
Police continue to investigate.