I am not a fan of Project Green Light (PGLD). For those unfamiliar with the program, according to their website, “Project Green Light Detroit is the first public-private-community partnership of its kind, blending a mix of real-time crime-fighting and community policing aimed at improving neighborhood safety, promoting the revitalization and growth of local businesses, and strengthening DPD’s efforts to deter, identify, and solve crime.” I’m not impressed.
I’ve previously laid out the problems with the program here, and here and here and here . I’ve blogged about it on the ACLU website and Channel 7 Did a story on it. I even sent a letter to every member of Detroit City Council laying out my issues. Yeah, I really don’t like Project Green Light.
The only Council Members to respond were Roy McCalister and Council President Brenda Jones. Recently, Chief Craig answered a number of questions posed to him by Council President: greenlight responses dpd. The answers are enlightening. Below you will find Council President Jones’ question, Chief Craig’s answer, and my comment on it. If you want a summary, here it is: There has been little research done on the effectiveness of Project Green Light and there are few policies in place to ensure it is not abused. But we already knew that.
Question 1. Do Green Light Locations have less crime than other parts of the City?
Response: While there are some parts of the city that have less crime than Project Green Light Detroit locations, Project Green Light Detroit partners have had larger decreases in violent crime than the city as a whole. At or within 250 feet of a Project Green Light Detroit locations, violent crime has reduced by 11% from the time they became partners in 2017 to December 31, 2017 as compare to the same period in 2016. Overall, the city has experienced a 7% decrease in violent crimes in 2017 as compared to 2016.
Comment: Chief Craig is trying to say PGLD has reduced violent crime by 4%, but this data doesn’t show that; it is statistical nonsense. Comparing the average drop in violent crime at PGL sites to the average drop in violent crime city-wide is apples and oranges. I could get all nerdy on you, but I won’t. Instead, just consider that this doesn’t show Project Green Light works. It’s an average. That 4% may well be the result of a huge drop at a few sites off-setting increases at others. And why the 250 foot radius? The impact of PGLD can only be determined by comparing crime rate changes at PGLD sites to those at comparable/nearby non-PGLD sites. Anything else is pure distraction.
Question 2. Has camera surveillance decreased crime as opposed to improved lighting or increased police presence that come with PGL?
Response: PGLD is a layered approach to reducing violent crime through camera surveillance, police presence, and adequate lights and signage. It would be a violation of the MOU for the business to not install and maintain the cameras and provide the adequate lights and signage. If the business fails to adhere to these policies, the partnership will terminate and they will no longer be a PGLD business. In turn, DPD can monitor these locations and provide patrol visits to the business.
Comment: This questions dives deeper into the issue raised by the prior question, but the answer completely dodges the basic question: how much of any crime drop can be attributed to the lights and increased police presence and how much can be attributed to the camera surveillance. This is an important question because the cameras and internet connection are the expensive (and potentially unconstitutional) parts of the program. Given that there is substantial evidence that live camera surveillance systems do not decrease crime – but increased lighting and police patrols do – the burden is on DPD to establish PGLD camera surveillance is as an exception before expanding the program. So, does Chief Craig not understand the question or not know the answer?
Question 3. Is there any evidence showing crimes being prevented due to live surveillance?
Response: As part of the Smart Policing Initiative (SPI) grant received by DPD from the Department of Justice, Michigan State University will evaluate a number of different factors of the program, including deterrence.
Comment: I guess this answers my last question. DPD doesn’t know if PGLD works. Yes, that’s right, Chief Craig admits DPD has no evidence the live-surveillance works! It’s nice they are having research done, even if plenty already exist. Here is the heart of the issue: If live monitoring doesn’t reduce crime, then the live surveillance is no better than requiring stores to have cameras and lighting. That video could be requested or sought with a warrant. From what we know right now, the money spent by businesses to stream live video and spent by the police to monitor that video is being wasted. At the very least, DPD should wait until they have supporting evidence of efficacy before expanding the program. There is no justification at all for considering an ordinance to make it mandatory!
Question 4. What regulations are there governing the access and distribution of video collected by PGL cameras?
Response: As stated in the MOU, DPD shall have live access to the video feed, and the business owner must maintain video on their local storage drive for up to 30 days. DPD may distribute the video to the public for the purposes of assisting an investigation of a criminal incident. State or federal authorities can have access to video only for evidentiary purposes pertaining to a particular crime incident.
Comment: The flaws here are what’s not said. You should take a look at Guidelines for Public Video Surveillance, by the Constitution Project. (I sent a copy to Council President Jones.) It lays out the principles for an effective and constitutional surveillance system. PGLD ignores all of the principles, including these:
- Under most circumstances, individuals may be “tracked” or “identified” by a public video surveillance system only pursuant to a warrant.
- Law enforcement must obtain administrative approval for secondary use of “pre-archival” stored video surveillance footage.
- Prohibit, to the extent possible, sharing of public video surveillance data with third parties, including private litigants, and restrict sharing with other governmental entities.
- Establish mechanisms to protect the rights of identifiable individuals captured on video surveillance data.
Question 5. How are equipment and service vendors selected?
Response: Businesses choose approved vendors to install the necessary equipment. Approved vendors must adhere to the requirements identified in the “Tier One Installer” and the “Tier Two Installer” agreements. These agreements can be found at greenlightdetroit.org.
Comment: Huh? Ok, I guess the question was phrased poorly. What I want to know is how do you get to be an approved vendor. The website doesn’t answer that question and neither does Chief Craig.
Question 6. What is the PGL or DPD policy on inclusion of individuals in a facial recognition database?
Response: DPD does not maintain a facial recognition database.
Comment: That’s good, but that wasn’t the question. Facial recognition technology works by capturing facial features and comparing them to those in a database. So, for this to work there has to be a database somewhere. I want to know who ends up in it and why. This is serious. You should read up on this technology, including it’s difficulty in recognizing minorities (Read here for the scholarly article or here for the NY Times summary) and the over-representation of minorities in the database used by law enforcement.
UPDATE: Project Green Light is now in Detroit schools.
Question 7 relates to policing certain communities. The answer given by DPD is disturbing, particularly in light of DPD’s recent history of routinely violating the civil rights of Detroiters. That is a post unto itself. Stay tuned.