Adapted from Veronica Beijos
The Boston Police Department maintains a vast network of surveillance cameras in public spaces throughout the city. According to the BPD, as of July 2024, the department oversaw 1,153 surveillance cameras, with an additional 150 cameras in and around BPD property. On top of these 1,300+ cameras, Boston Police also have access to footage from approximately 400 cameras owned by other city agencies, like the Boston Transportation Department. This surveillance camera network has expanded dramatically in the last few years, with 395 more cameras added since 2020, amounting to an increase of 50%. (The MBTA also has thousands of its own cameras; it’s not clear whether or how they are linked to the BPD’s surveillance network. The discussion below does not include MBTA cameras.)
These cameras are part of a larger network operated by the cities in the Metro Boston Homeland Security Region, which is funded by the Department of Homeland Security. Historically, the public has known little about how the department uses this technology, or what civil rights and civil liberties protections might be in place to guard against misuse and abuse.
For instance, we know that the BPD uses machine learning technology called “BriefCam” to automate the analysis of surveillance camera footage, enabling officers to track objects and people across BPD’s entire camera network. But we know virtually nothing about how the BPD uses this powerful surveillance technology, even though it raises a host of civil rights and civil liberties issues. Vast camera networks supercharged by powerful software can chill speech, enable police to track protesters and dissidents, and facilitate dragnet surveillance of ordinary people going about their lives, even if those people aren’t doing anything unlawful or suspicious.
As we have shown elsewhere, surveillance is not spread evenly across the city. With surveillance cameras, the story is much the same. The data shows that the downtown business districts contain the most cameras. But outside of downtown, in the largely residential neighborhoods of the city, cameras are disproportionately concentrated in communities of color. For example, West Roxbury, a mostly white community, has very few cameras, while the much more racially diverse neighborhoods of Dorchester, Roxbury and the South End have over 100 each.
A BPD surveillance camera in downtown Boston. (Photo by Veronica Beijos)
In 2020, the Boston City Council unanimously passed a surveillance oversight ordinance, in part to understand whether there is any daylight between how police justify surveillance technologies and how they use those technologies. Among other mandates, the ordinance requires the city to publish annual surveillance reports detailing how approved surveillance technologies were used during the prior year.
Last year, in compliance with this ordinance, the BPD published a document showing every time a police officer requested surveillance video from the department's Video Evidence Unit, including the district and unit that requested it and the type of incident under investigation.
Unfortunately, the document in question is missing a key piece of the puzzle: namely, whether the surveillance video was ultimately useful to investigators. That is, the data does not mention whether the footage in fact captured an alleged crime, whether said footage was then used to charge someone or whether the footage was instrumental in convicting them.
But while the data doesn’t tell us everything, it does provide a first ever glimpse into how the BPD uses the 1,300+ cameras at its disposal.
This chart shows the incidents for which BPD officers requested surveillance video in 2023, grouped by incident type. Categories with very few requests were omitted for visibility.
By analyzing the video request log, we found that the most common reason for requesting surveillance video in 2023 was to investigate motor vehicle accidents.
In fact, MVAs made up around 1 in 5 of all police surveillance video requests.
You can explore more details about requests for surveillance video footage broken down by unit below:
Police usually justify the use of surveillance cameras with two arguments: first, that surveillance cameras help deter crimes; and second, that surveillance cameras help solve crimes. But they rarely provide evidence for these claims.
When it comes to the question of deterrence, a meta-analysis of 24 studies over 40 years found that surveillance cameras did not significantly reduce crime in the US. But even if cameras don’t deter crimes, police often point to a different argument: surveillance footage helps them solve crimes after the fact. Does the data support this claim?
In short, no. In London, one of the most heavily surveilled cities in the world, an internal report found that police only solved one crime for every 1,000 cameras in the city. In Boston, we have known relatively little about how police use their vast surveillance camera network — until recently.
The first of the DHS-funded surveillance cameras in Boston were installed in 2004. Since then, the BPD and other police departments in the Greater Boston area have vastly expanded their surveillance toolkit. While the police say surveillance technologies like cameras are an integral part of their investigations and vital to solving serious crimes, we wanted to test this claim. So, we started with the most serious of crimes: homicides.
While Boston has one of the lowest murder rates in the country, to our knowledge, there is no evidence that surveillance cameras have played a part in that success story. But have cameras been a game changer in solving homicides after the fact? Again, the BPD did not provide us with enough information to answer this question conclusively. However, we can look at homicide clearance rates and the growth of surveillance in Boston to see if there is any correlation.
The data shows that, despite the explosive growth of camera surveillance in Boston over the past twenty years, homicides often remain unsolved, and there hasn’t been a dramatic improvement in the proportion of solved (“cleared”) cases since the introduction of surveillance cameras in 2004.
In 2003, the year before the first of the DHS-funded cameras were installed, the police cleared more than half of all homicides in the city. But 1,300+ cameras later, the clearance rate is not much higher than when it started, at 58% cleared in 2024.
In fact, Boston's clearance rate is consistently lower than the national average, despite a concurrent massive expansion of the BPD’s surveillance ecosystem, which includes not just surveillance cameras but also license plate readers, the “ShotSpotter” gunshot detection system, and the Boston Regional Intelligence Center (BRIC).
Per the ordinance, Boston must publish its annual surveillance report every year by May 31. Last year, the city did not release its report until July — and this year it is also late to report. We will update this page when the latest report becomes available.
In the meantime, the BPD should share more information about its camera surveillance with the public. While the findings laid out above are interesting, absent more detailed information about how the footage was used, Boston residents and city councilors overseeing the BPD’s surveillance programs can’t gauge whether these cameras are worth the civil liberties infringements or financial cost. We need more data — specifically, when footage was ultimately used to charge or convict someone, and whether it was instrumental in this process. The number of times that camera footage was useful to police is likely lower, perhaps far lower, than the total number of requests.
The surveillance ecosystem in Boston keeps growing, too often without evidence-based justification. Though police do share anecdotes about the utility of surveillance cameras, we do not have comprehensive data to support the broader claim that surveillance increases safety. And the data we do have indicates that cameras are not the miracle crime solving technology that CSI and other television shows would have us believe.
Here’s what we do know: coupled with artificial intelligence enabling automated analysis, surveillance camera networks pose serious threats to our basic civil liberties. While Boston City Council recognized the harms and pitfalls of facial recognition in 2020 — and banned its use — few Bostonians are aware that similar technologies enabling dystopian tracking are still operational. With software like BriefCam, police can automatically track people, vehicles and other objects across multiple cameras, essentially enabling real-time and historical mass surveillance. Despite the grave risks posed by such technologies, the BPD provided zero information to the public in last year’s surveillance report about its use of BriefCam.
The surveillance oversight ordinance was designed to inform residents and policymakers about the actual impacts of surveillance technologies on our communities. While we have made significant progress towards that end, we still have work to do. Join us in calling on the BPD to make more detailed information available so we can judge their claims about surveillance efficacy against the actual data.
Julie J Lee led data analysis, data visualizations and writing, with thanks to Kade Crockford for comments. Renee Perpignan developed the front-end for the initial implementation of the scrolling visualization.