Police Body Worn Cameras in Rio’s Favelas: Can Technology Reduce Violence?

In 2016, a team of three researchers based at Stanford University —Beatriz Magaloni, Vanessa Melo, and Gustavo Robles— conducted a groundbreaking experiment in Rocinha, Rio de Janeiro’s largest favela (informal settlement), to test whether body-worn cameras (BWC) could reduce police violence and improve community relations.

The findings reveal that body cameras hold great promise, but they also come with serious challenges.  Before the experiment started, one police unit commander ominously told the researchers: “If you give body cameras to my officers, this will stop them from doing their job.”

Favela settlements in Rio De Janeiro are known for the high levels of poverty and the presence of organized criminal groups that, for years, have served as “rulers” of these territories. Rocinha (meaning ‘little farm’ in Portuguese) has 120,000 residents who are now policed by the UPP (Pacifying Police Unit).

Rio de Janeiro’s militarized police force is among the deadliest in the world. Between 2003 and 2019, officers killed nearly 20,000 people, representing about 20% of all homicides in the region. The police officially justify these killings as ‘auto de resistência’ (resistance to arrest), and such killings are rarely investigated by the judiciary.  

In Rocinha, the relationship between the police and the community was found to be deeply concerning. A survey before the study found that residents were more likely to be victimized by police than by criminal groups; indeed, 61% of residents reported fearing being killed by police, and only 10% said they would call the police to resolve a conflict. Common complaints included officers slapping suspects, pulling hair, using excessive force, and making threats.

The violence went both ways. Police officers reported being cursed at, having water, urine, or stones thrown at them, and suffering verbal threats and physical attacks from residents. This mutual hostility led to a vicious cycle of mutual aggression and mistrust.

From December 2015 to November 2016, the research team randomly assigned body cameras to police shifts in various units within Rocinha’s UPP. The study tracked 857 different officers across 52,000 shifts, making it the first randomized controlled trial of police body cameras in a developing world setting.

Officers were required to activate their cameras during all interactions with civilians. The goal was to see if cameras would reduce aggressive police behaviors, decrease the unnecessary use of force (especially deadly force), improve police reaction times, and finally reduce community aggression towards officers.

The researchers’ findings were complex. On the one hand, they discovered that camera assignment led to a 39% reduction in stop-and-search. In Rocinha, reduced proactive enforcement may be a positive shift, as residents reported frequent police abuse in everyday interactions with officers.

The researchers also found suggestive evidence that BWCs reduced the use of deadly force. During the study, police fired their weapons in 27 incidents, discharging a total of 469 bullets—and 70% of these were fired when officers were not wearing cameras. Even more striking, the total number of shots fired in Rocinha dropped sharply compared to the previous year’s staggering 3,666 rounds. While these results point to the potential of body cameras to curb deadly force, the researchers caution that this conclusion must be treated with extreme care, given the relatively small number of shooting incidents observed.

Fascinatingly, community aggression toward the police also decreased during the study. Reports of residents throwing objects at officers or making threats declined significantly, suggesting the deterrent effect worked both ways — civilians may have moderated their behavior after realizing they were being recorded.

However, to their surprise, the team discovered BWCs also produced a strong “de-policing” effect: officers wearing cameras were 43% less likely to respond to emergency calls from the operations center, and 60% less likely to respond to direct requests for help from residents on the street. This showed that while officers did reduce violent interactions with favela residents, they also stopped performing basic necessary police work.

The Compliance Problem

One of the most troubling findings was that officers frequently disobeyed the camera protocols, refusing to turn on their cameras during their shifts. 50% of reported police–civilian interactions were never recorded. Moreover, camera usage significantly dropped throughout the year-long study, from 40% compliance early on to less than 5% by August 2016.

The research revealed a troubling pattern: the officers most likely to have injured civilians in the past were also the most resistant to recording their interactions. Put simply, those most in need of oversight were the least likely to comply. In contrast, officers who did activate their cameras were often those who had experienced more aggression from the community—suggesting that some saw the footage as a tool to deter such confrontations and protect themselves.

The Limitations of BWCs in Rocinha

The study revealed deep cultural resistance to BWCs within the Rocinha UPP. One anonymous officer told the researchers that there was a “duality inside the organization. They teach us one thing but expect something different.” He reported that while officers receive messaging about being more community-focused, they are still trained like soldiers to wage “a war against crime”.

Many officers also expressed the belief that cameras would prevent them from doing their jobs effectively. They worried about creating evidence that could be used against them, with one saying, “Nobody is obligated to generate proof against themselves… but that happens with the police officer when he is wearing a camera.”

Officers described feeling abandoned by the state in “enemy territory”. Working from a small metal building riddled with bullet holes, they saw themselves as “warriors” in hostile territory rather than community servants. This siege mentality made them resistant to any intervention that might limit their discretion or ability to use force.

While officers on the frontline were reluctant to wear cameras, the problem extended further into the upper echelons of command. Police supervisors failed to enforce camera protocols. Only 35% of officers reported being supervised regarding camera usage, and just 9% received warnings for non-compliance. The unit commanders themselves believed cameras would prevent officers from doing their jobs, undermining any effort at enforcement.

Importantly, when supervisors were randomly assigned cameras, officers under their command showed a 300% increase in policing activities – suggesting that monitoring supervisors, not just frontline officers, may be crucial.

This study shows that BWCs alone cannot dismantle deeply entrenched cultures of violence and impunity. In contexts where police view themselves as warriors rather than guardians, where supervision is weak, and where relations with the community are severely damaged, technological solutions must be paired with far-reaching institutional reforms to make a meaningful impact.

The researchers suggest several important lessons. The first is that BWCs require strong institutional support. Without strong supervision, many officers will refuse to turn their cameras on when interacting with civilians. Second is that assigning cameras to supervisors has a strong positive effect on frontline officers.

While body cameras show promise in reducing aggressive encounters between police and civilians, as well as aggression toward officers, their full potential has yet to be realized. Challenges remain—such as “depolicing” and inconsistent compliance, with some officers failing to activate their cameras during civilian interactions. In contexts of extreme violence and mutual distrust, these obstacles highlight that technological tools work best when embedded in broader institutional and cultural reforms.

One key limitation is the discretion officers currently have in turning cameras on and off. In settings where protocol violations are frequent, remote activation from the central station—removing this decision from frontline officers—could ensure more consistent use. This technology already exists and could be particularly valuable in high-conflict areas like Rio de Janeiro.

For body cameras to deliver their intended impact, police departments need leadership committed to accountability, supported by systems that monitor not only officers but also their supervisors—potentially by assigning cameras to them as well.

The Rocinha experiment offers important lessons: body cameras are a powerful tool, but not a “magic bullet” for police reform. Cameras can contribute to a more transparent, trusted, and accountable policing system only with the right safeguards, leadership, and a parallel effort to address the institutional, cultural, and social conditions that fuel police violence.