By The Global Initiative Against Transnational Organized Crime

 Years of escalating gang violence in the Western Cape has sustained the worrying upward trend of the past five years. This issue of the Western Cape Gang Monitor summarizes the factors that have driven gang dynamics in 2025 and sets out a plan to tackle the challenge in the short term. 

 As we move into 2026, this plan can help form a basis for decisive action against escalating gang violence – and support for the individuals and communities it endangers and harms.

In this issue:

• Gang dynamics: 10 trends.

• What generates clusters of violence?

• Know your enemy: the ever-shifting challenge.

• A 12-point plan for the rapid mitigation of gang violence.

This is the seventh issue of the Global Initiative Against Transnational Organized Crime’s (GI-TOC) Western Cape Gang Monitor, an output of our South Africa Organized Crime Observatory. This series of bulletins tracks developments in Western Cape gang dynamics each quarter, to provide a concise synthesis of relevant trends to inform policymakers and civil society. This is a year-in-review issue, combining analysis published by the GI-TOC throughout the year with new research. The monitor draws on information provided by field researchers working in gang-affected communities of the Western Cape. This includes interviews with current and former gang members, civil society and members of the criminal justice system.

By Sou Lee and John Leverso 

Positionality is an important consideration when carrying out research. An effective tool for understanding this process is reflexivity—a continual dialogue that explores the interplay between our identities and how data is collected, analyzed, and interpreted. These reflexive accounts have been used in various disciplines, including criminal justice and criminology. In advancing this important practice, we offer insight into our experiences studying a hard-to-reach population: gangs. Specifically, we document how our insider and outsider identities, as well as the space between facilitated access, were used strategically and informed our interpretations of data. We conclude by encouraging reflexivity within criminology broadly and specifically among scholars who study hard-to-reach populations like street gangs.

Qualitative Criminology (QC): Vol. 13: No. 2, Article 4. 2024 

By Peter Honnef

On paper, gang databases look efficient and innocuous, providing law enforcement agencies with a shortcut to identify people suspected of having a connection to gang violence. But these local databases often contain inaccurate information that, when shared widely across agencies, can have life-altering consequences. The Center for Policing Equity’s (CPE) new white paper, Gang Databases and Immigration Enforcement, explores the disproportionate impact gang databases have on marginalized communities and the increased harm from immigration enforcement. By Peter Honnef

On paper, gang databases look efficient and innocuous, providing law enforcement agencies with a shortcut to identify people suspected of having a connection to gang violence. But these local databases often contain inaccurate information that, when shared widely across agencies, can have life-altering consequences. The Center for Policing Equity’s (CPE) new white paper, Gang Databases and Immigration Enforcement, explores the disproportionate impact gang databases have on marginalized communities and the increased harm from immigration enforcement. 

Center for Policing Equity, 2026. 11p.

By Alison Siegler and Grant Delaune

Although the U.S. Sentencing Commission’s drug guidelines were intended to align punishment with culpability, decades of sentencing practice reveal a different reality. Outcomes are primarily driven by drug type and quantity, which have proved to be poor proxies for assessing culpability. We put forward a new approach to drug sentencing that instead focuses on a person’s function in a drug enterprise. We also propose anchoring base offense levels to pre-Guidelines sentencing data. To illustrate this model, we present a rewritten version of Guideline § 2D1.1 incorporating our proposed framework. Focusing on function rather than drug type and quantity will ensure that sentences are calibrated based on culpability and better fulfill the core purposes of punishment.

University of Chicago Law School, Public Law & Legal Theory Research Paper No. 25-38,  2025,

Editors: John K. Roman, Philip Cook

One of the great policy successes of the last decade is the increasing role of rigorous, objective, and transparent data and research in policymaking. Developing and implementing a data-driven government in which valid and reliable evidence informs solutions to our nation’s most pressing health and safety challenges is more critical than ever as those challenges are ever more complex. Nowhere is that data foundation more needed than in the realm of firearms violence. Trustworthy data is a much-needed bridge to effective policymaking that can reduce the number of firearm accidents, suicides, homicides, and assaults. In an age of intense partisanship, shared facts are the cornerstone for building a shared purpose. The shared purpose of modernizing firearms data infrastructure is to improve public safety by reducing gun violence. In the fall of 2020, Arnold Ventures, a philanthropy dedicated to maximizing opportunity and minimizing injustice, and NORC at University of Chicago, an objective nonpartisan research institution, released the Blueprint for a US Firearms Infrastructure (Roman, 2020)1 . The Blueprint is the consensus report of an expert panel of distinguished academics, trailblazing practitioners, and government leaders. It describes 17 critical reforms required to modernize how data about firearms violence of all types (intentional, accidental, and self-inflicted) are collected, integrated and disseminated. This project, which is also supported by Arnold Ventures, takes the conceptual priorities described in the Blueprint and proposes specific new steps for implementation. The first step in building a better firearms data infrastructure is to acknowledge where we currently stand. In The State of Firearm Data in 2019 (Roman, 2019)2 , the expert panel found that while there are a substantial number of data sources that collect data on firearms violence, existing datasets and data collections are limited, particularly around intentional injuries. There is some surveillance data, but health data on firearms injuries are kept separately from data on crimes, and there are few straightforward ways to link those data. Data that provide context for a shooting—where the event took place, and what the relationship was between victim and shooter—are not available alongside data on the nature of injuries. Valuable data collections have been discontinued, data are restricted by policy, important data are not collected, data are often difficult to access, and contemporary data are often not released in a timely fashion or not available outside of specialized settings. As a result, researchers face vast gaps in knowledge and are unable to leverage existing data to build the evidence base necessary to adequately answer key policy questions and inform firearms policymaking.

By The National Institute for Criminal Justice Reform

The Gun Violence Reduction Strategy (GVRS) is a comprehensive method of identifying the individuals in a jurisdiction who are at the very highest risk of being involved in gun violence and employing effective interventions to intervene with those individuals and their conflicts to significantly reduce gun violence. The National Institute for Criminal Justice Reform (NICJR) works with jurisdictions to implement GVRS through data analysis, policy development, training, coaching, management support and technical assistance. NICJR helps identify people and groups at highest risk of being involved in gun violence in a specific city; trains community leaders and law enforcement to effectively communicate that risk to those individuals in a supportive manner; helps stand up, manage, and train staff for intensive community violence intervention (CVI) efforts to work with those individuals; and works with law enforcement to focus their efforts in a way that reduces their overall footprint in the community and increases their focus on the small number of people driving gun violence. When operated to scale, well managed, and coordinated, GVRS can significantly and consistently reduce gun violence citywide. GVRS is a distinct strategy that has evolved from similar models known as Ceasefire and Gun Violence Intervention. ¹ GVRS rests on two main principles: specificity and intensity: Specifically serving those at the very highest risk of being involved in gun violence soon, and intensively and frequently engaging those at the very highest risk.

By Alon Steinberg and Amanda Frugoli 

Sudden and unexpected arrest-related deaths are deeply tragic and have generated widespread concern among the public, medical professionals, and law enforcement agencies. One mechanism that has garnered considerable attention is the use of prone restraint, wherein a subject is placed face-down and controlled in this position. The safety and risks of this technique remain subjects of debate within both scientific literature and legal settings. Supporters of prone restraint’s safety frequently cite prospective epidemiologic studies that report no fatalities associated with its use. However, these studies typically involve small cohorts and are conducted over limited timeframes, potentially underestimating rare but serious outcomes. In contrast, retrospective analyses, which assess larger populations over extended periods, have identified multiple cases of fatal outcomes linked to prone restraint. Notably, some of the most comprehensive data on these fatalities come from investigative journalism, which has uncovered patterns and prevalence rates not fully captured in academic or institutional studies. Based on available evidence, we estimate the mortality rate with use of in-custody prone restraint is at approximately 1 per 4.4 million individuals per year, or 0.023 per 100,000 population annually. These findings underscore the need for more rigorous, large-scale, and transparent epidemiological studies to better inform public policy, law enforcement practices, and clinical guidelines. The potential lethality of prone restraint must be recognized, and its use re-evaluated in light of both fatal risk and ethical responsibility.

By Justin M. Feldman, Tracey Lloyd , Phillip Atiba Solomon

 Background: Mortality classification for deaths in US police custody has important consequences for epidemiologic monitoring and legal outcomes. Prior literature suggests in custody death classification is inconsistent and may not reflect non-firearm force that preceded death. Methods: We analyzed the Associated Press “Lethal Restraint” national dataset (United States, 2012-2021; N = 1,036), which included deaths following police use of non-firearm force. Our primary outcomes included whether the death investigator: (1) classified manner of death as a homicide, (2) mentioned a force-related injury/condition in the cause-of-death statement, and (3) mentioned any force. Inverse-probability-weighted logistic models estimated the association of these outcomes with death-investigator jurisdiction type, local political composition (quartile of Republican Party vote %), decedent race/ethnicity, and each agency’s prior classifications. Findings: We removed 96 deaths based on exclusion criteria. Of the remaining 940 deaths, 28.5% were classified as homicide, 16.5% had cause-of-death statements mentioning a force-related injury/condition, and 42.6% mentioned any force. In contrast, 73.9% of statements mentioned drugs. Unadjusted results showed homicide classification increased from 25.0% in 2012-2014 to 32.2% in 2018-2021. Models estimating adjusted prevalence differences (aPD) showed that, compared to medical examiner jurisdictions, coroners (aPD: -0.19; 95% CI: -0.31, - 0.06) and sheriff-coroners (a PD: -0.17; 95% CI: -0.28, -0.05) were less likely to classify deaths as homicides. Model results also showed that classifications for incidents occurring in the leastRepublican counties were most likely to reflect force across all three manner and cause outcomes. Interpretation: Non-homicide classifications and cause-of-death statements making no mention of force were widespread for US in-custody deaths. We identified novel evidence suggesting coroner and sheriff-coroner jurisdictions were especially unlikely to categorize in-custody deaths as homicides, and that incidents occurring in highly Republican counties were least likely to reflect force in the cause or manner of death. 

By Victor Weedn , Alon Steinberg , Pete Speth 

We postulate that most atraumatic deaths during police restraint of subjects in the prone position are due to prone restraint cardiac arrest (PRCA), rather than from restraint asphyxia or a stress‐induced cardiac condition, such as excited delirium. The prone position restricts ventilation and diminishes pulmonary perfusion. In the setting of a police encounter, metabolic demand will be high from anxiety, stress, excitement, physical struggle, and/or stimulant drugs, leading to metabolic acidosis and requiring significant hyperventilation. Although oxygen levels may be maintained, prolonged restraint in the prone position may result in an inability to adequately blow off CO2, causing blood pCO2 levels to rise rapidly. The uncompensated metabolic acidosis (low pH) will eventually result in loss of myocyte contractility. The initial electrocardiogram rhythm will generally be either pulseless electrical activity (PEA) or asystole, indicating a noncardiac etiology, more consistent with PRCA and inconsistent with a primary role of any underlying cardiac pathology or stress‐induced cardiac etiology. We point to two animal models: in one model rats unable to breathe deeply due to an external restraint die when their metabolic demand is increased, and in the other model, pressure on the chest of rats results in decreased venous return and cardiac arrest rather than death from asphyxia. We present two cases of subjects restrained in the prone position who went into cardiac arrest and had low pHs and initial PEA cardiac rhythms. Our cases demonstrate the danger of prone restraint and serve as examples of PRCA.

By ALISSA SKOG AND JOHANNA LACOE 

Research series examining second look policies in California The five policy briefs and overview report in this series describe the characteristics and recidivism rates of individuals affected by second look policies in California. Committee on Revision of the Penal Code Before 2019, people who participated in certain felonies that resulted in a death could be convicted of murder, even if the person had neither committed the killing nor intended for it to occur. These convictions were based under the felony murder rule or the “natural and probable consequences” doctrine, which both allowed for broad liability. To address concerns about fairness and excessive punishment, Senate Bill 1437 (2018) narrowed or eliminated the application of these doctrines and allowed individuals convicted under them to petition for resentencing. In recognition that other serious convictions were similarly affected, Senate Bill 775 (2021) expanded eligibility to include those convicted of manslaughter or attempted murder under these legal theories. This brief examines who was resentenced under felony murder reform, the offenses for which they were originally convicted, and their recidivism rates following release. Key findings • Nearly 1,200 people were resentenced after these changes to the felony murder rule. As of December 2024, 1,172 people initially convicted under the felony murder rule have had those charges vacated, and were resentenced based on the remaining charges in their cases. Of those resentenced, 78% have since been released from prison. • Most people convicted of felony murder were young at the time of the offense, and for many, it was their first admission to prison. The median age at the time of the offense was just over 21, and for 75% of those resentenced, that conviction was their only prison sentence. • Women made up a larger share of those resentenced under felony murder reform (9%) than the share of all people released from prison in fiscal year 2018–19 (7%). • Recidivism rates for those resentenced under felony murder reform (9%) than the share of all people released from prison in fiscal year 2018–19 (7%). • Recidivism rates for those resentenced under felony murder reform were notably low. New conviction rates were consistently lower than the total releases (3% within one year, 7% within two years, and 10% within three years, compared to 21%, 33%, and 42%). Of those resentenced and released, most new convictions were misdemeanors. • Very few people were convicted of a new serious or violent felony after resentencing under felony murder reform. Fewer than five people were convicted within one- and two- years, and only 2% (n=5) were convicted within three years. However, we can only assess full three-year outcomes for 25% of people who have been released (n=274).

By William Downs

Who handles complaints against the police?

A member of the public can make a complaint if they are dissatisfied with the police. 

There are three crucial actors in the police complaints system:

• Professional standards departments (PSDs) are specialist teams based within every police force in England and Wales. They are responsible for handling most complaints for their force.

• The Independent Office for Police Conduct (IOPC) is an independent body that oversees the police complaints system. It also conducts independent investigations into some of the most serious police complaints and conduct matters.

• Local policing bodies (either the police and crime commissioner or the deputy mayor for policing and crime, depending on the area) are responsible for monitoring their force’s complaint handling and conducting some complaint reviews.

By Anna Cook, Jon Lloyd, Fablina Sharara, Jennifer Key,

When someone is in a crisis, a police response can lead to help — or harm. Across the country, communities are trying something new: sending mental health specialists, peer specialists or other trained professionals to crisis calls instead of police through alternative response programs (ARPs). Research on ARPs has focused on case studies and standout programs like CAHOOTS in Oregon and Denver STAR, but we know relatively little about the broader trends in this growing field. Comprehensive information about where and how jurisdictions are implementing ARPs is crucial so that policymakers, funders and advocates can make better informed strategic decisions regarding public safety innovation. To address this gap, we created a novel database of 216 ARPs established since the early 1970s and operational as of 2024 to produce one of the first overviews of these programs throughout the United States. By summarizing the design, scale and geographic distribution of ARPs, we provide a broad look at the field to help inform and empower community leaders to build stronger public safety systems while reducing dependence on traditional policing. Our findings underscore critical choices in how ARPs are implemented and raise important questions about their scope, accessibility and long-term potential. As policymakers, practitioners and advocates continue to explore alternatives to police response, this report provides a foundation for understanding the current landscape and identifying paths for growth. Given current gaps in ARP implementation, future research and innovation are needed to explore how these programs can evolve to handle a higher volume and wider range of calls, understand the benefits and limitations of different call lines, and expand to meet the needs of smaller or underserved communities.

 Key findings: • Recent proliferation: Public officials and other decisionmakers established nearly 120 ARPs from 2020 through 2024, reflecting a surge in interest and political willpower following national Black Lives Matter protests. • Limited scale: Most programs serving large populations respond to fewer than five calls per 1,000 residents per year. • Narrow scope: Mental and behavioral health are a stated focus for 94% of programs; far fewer are designed to address issues like traffic safety, interpersonal conflict or homelessness. • 911 reliance: Despite their focus on mental crisis calls, only 18% of ARPs use the 988 national mental health crisis line, while 50% use 911 for dispatch. • Urban concentration: Programs are concentrated in large, racially diverse, urban areas.

West Hollywood, CA,

WeWest Hollywood, CA,West Hollywood, VThe Center for Policing Equity’s (CPE) , 2025. 23p.

Assessing Multi-disciplinary Response Mechanisms

Edited by Paromita Chattoraj

Child Protection in India: Assessing Multi-disciplinary Response Mechanisms offers a comprehensive exploration of the institutional, legal, and social frameworks surrounding child protection in India. Anchored in a multidisciplinary approach, the book brings together insights from law, social work, psychology, education, and public policy to examine how various systems interact in addressing the issues related to protection of children from abuse, neglect, trafficking, and exploitation. This book is intended for researchers, academicians, legal professionals, social workers, policymakers, child rights activists, and students engaged in child welfare and protection studies. It critically analyses existing response mechanisms by stakeholders such as the Juvenile Justice Boards, Child Welfare Committees, Child Care Institutions, police, and the judiciary, and the implementation of various laws, including Juvenile Justice (Care and Protection) Act, the POCSO Act, Prohibition of Child Marriage Act, while also highlighting challenges in inter-agency coordination. By combining chapters under four broad themes focusing on empirical research, policy review, and case studies, this volume equips readers with a nuanced understanding of the gaps and strengths in current practices and highlights of best practices around the world. The book aims to foster a rights-based and child-centric approach, encouraging collaboration across disciplines and sectors. It is a valuable resource for anyone seeking to engage meaningfully with child protection frameworks in India and to contribute to more effective, sustainable interventions.

London: Routledge, 2026. 

By Cornelius Friesendorf

For post-conflict stabilisation and peacebuilding efforts to have a chance of success, security gaps need to be closed. Domestic law enforcement agencies tend to be too weak or unreliable after war to enforce the law and fight serious crime. As a consequence, operations against organised crime, the arrest of suspected war criminals and the protection of minorities depend on international intervention forces. Much attention has been paid to domestic police reform and the problems of deploying international civilian police. This book examines the under-explored role of international military missions in post-conflict law enforcement, with a focus on serious crime. The military is under pressure to fill security gaps. Yet military involvement in crime-fighting is problematic practically (soldiers are generally not trained and equipped for it), politically (crime-fighting is seen as military mission creep) and normatively (it undermines the delineation of military and policing functions). Military support of law enforcement poses a major dilemma in peace operations. Decision-makers continue to struggle with this dilemma in an ad-hoc fashion, while scholars have so far provided few empirical accounts. This book focuses on post-war Bosnia-Herzegovina (BiH) and Kosovo. It argues that the pros of military involvement in law enforcement outweigh the cons, given the continuing lack of strong police forces, the criminalised nature of contemporary wars and the negative influence of spoilers onstabilisation and peacebuilding efforts. Military support of the fight against serious crime is not sufficient for these efforts to succeed, but it is necessary. This book has three objectives. It describes the role of NATO and EU military missions in law enforcement in BiH and Kosovo from the 1990s through early 2009, thus filling an empirical knowledge gap. By formulating a strategy for military involvement in law enforcement, the book also makes a normative contribution to the debate on peace operations. Drawing on, among other sources, interviews in these two ‘international protectorates’, the book shows that military support of the fight against serious crime has lacked effectiveness, efficiency and legitimacy, examples of success notwithstanding. Third, the book adds conceptual value to the debate on peace operations, by drawing on security governance, Security Sector Reform and Security Sector Governance. These concepts help to understand the military role in post-conflict law enforcement and to guide improved efforts.

Geneva: Geneva Centre for the Democratic Control of Armed Forces (DCAF) , 2010. 207p.

By Nicolas Florquin with Victor de Oliveira

SITUATION UPDATE: LATIN AMERICA

Based on UN Comtrade data, reported global small arms and light weapons (hereafter ‘small arms’) exports rose sharply from USD 5 billion in 2019 to USD 9.2 billion in 2024. Consistent with previous trade updates, ammunition remains the most traded weapon category, accounting for 35% of the value of reported global imports for the period 2019–24, followed by sporting and hunting shotguns and rifles (21%), and pistols and revolvers (18%).

The Authorized Trade in Small Arms: Latin America from a Global Perspective—a Situation Update from the Mapping the Transnational Circulation and Control of Small Arms in Latin America project—examines the global authorized trade in small arms between 2019 and 2024, with a particular focus on trends in Latin America. It finds the region to be a comparatively small player in the global authorized small arms trade, accounting for 2.8% of global small arms imports and 6.3% of exports. Yet military firearms represented about 10% of all Latin American small arms imports during this period—almost double the global average of 5.6%.

The Situation Update also identifies a significant increase in European—and in particular Eastern European—imports during this period, which seem to have fuelled the growing trade. Indeed, European imports accounted for 40% of all reported global imports in 2024, while the value of Eastern European imports multiplied by more than ten between 2019 and 2024.

Geneva: Small Arms Survey, 2025. 16p.

By Alina D. Machande 

The burgeoning field of cyberdeviance lacks a unified conceptual framework, hindering classification and understanding of its subtypes and underlying psychological mechanisms. To address this gap, we conducted two studies. In Study 1 (N = 20), employing the repertory grid technique, we identified five key dimensions of cyberdeviance. In Study 2 (N = 268), participants rated 16 cyberdeviant behaviors on these dimensions, revealing three subtypes: data-oriented, interpersonal, and non-prototypical cyberdeviance. Our findings suggest a shift from singular cyberdeviance investigation toward recognition of its diverse subtypes, each necessitating tailored interventions. By adopting a dimensional approach, we transcend categorical and technocentric perspectives, enabling examination of behavior clusters across cultural and temporal contexts. Our work underscores the importance of integrating foundational deviance theories and expanding conceptual frameworks to comprehensively grasp cyberdeviance phenomena.

The Information Society, 1–19. 2025.

By Steven Schuetz, et al.

  Context Service weapon activity, including instances where an officer’s firearm is drawn, pointed, or discharged, plays an important role in understanding events transpiring during a police–public encounter. Detection, documentation, and communication of these events in a way that is accurate, timely, and dependable is vital for enhancing transparency and accountability of law enforcement service weapon use. About this Report The National Institute of Justice (NIJ) requested the Criminal Justice Technology Testing and Evaluation Center (CJTTEC) to investigate the landscape of commercially available and emerging technologies that could meet this need. CJTTEC conducted a review of technologies capable of detecting when a service weapon has been unholstered, pointed, or discharged; documenting when a law enforcement officer discharges their service weapon (or initiating documentation such as body-worn camera (BWC) recordings in such incidents); and communicating the information to dispatchers. CJTTEC’s methodology to understand this technology landscape included secondary research (e.g., reviewing patents, trade literature, press releases, news articles, and publications) and primary research with technology experts, product representatives, and researchers. This brief provides a high-level summary of technology systems capable of documenting, detecting, and communicating service weapon activity, focusing specifically on technology integrated into or onto the weapon, in a holster, in a BWC, in a wearable device, or in environmental sensing tools. Conclusion Although no single commercially available tool is capable of detecting, documenting, and communicating service weapon activity, law enforcement agencies may be able to rely on a suite of products to help them address these needs.  Key Takeaways ¡ Agencies are facing increased pressure to document service weapon activity. From 2015 through 2020, on average, an estimated 1,769 people were injured annually—979 fatally and 790 nonfatally— from shootings by police in the United States.1 Because of the impact that officer-involved shootings (OISs) have on the community, law enforcement agencies are facing increased public pressure and policy mandates to document service weapon activity. ¡ There is a need for tools or technologies that can objectively detect, document, and communicate service weapon activity. OISs are stressful incidents that can occur quickly and under poor visual circumstances, which can impact accurate documentation of events. Further, obtaining reliable service weapon activity data can be challenging because of noncompliance with body-worn camera (BWC) policies, lack of BWCs, or inaccurate witness and officer accounts. ¡ There is no single commercially available product that meets service weapon activity needs. No single product can currently (1) detect service weapon activity, such as recording actual shots fired in an incident involving law enforcement weapons; (2) document the activity, such as initiating BWC recordings; and (3) communicate information about service weapon activity to police dispatchers. ¡ Agencies can rely on a suite of products to address these needs or choose specific products, each with strengths and limitations. Available technologies may be integrated into or onto the weapon, in a holster, in a BWC, in a wearable device, or in environmental sensing tools. Weapon-integrated tools offer the most functionality to detect and document multiple types of service weapon activity during a use-of-force incident, but many of these products, such as those developed by Armaments Research Company and Yardarm, are not commercially available. These products often lack the capability to communicate updates in real-time with dispatch. Holster-integrated tools can sense officer unholstering activity, activate BWC, and communicate with dispatch, but they cannot detect activity related to pointing or firing a weapon. BWCs, activated by multiple types of triggers, can document audio and video of the incident and communicate with dispatch, but they cannot specifically detect officer firearm activities (e.g., weapon unholstering, pointing, gunshot detection). Wearable devices can detect officer firearm activities, document metadata, and communicate with dispatch, but most products are still in a development phase for law enforcement applications. Environmental sensing tools may detect and document activities transpiring within a certain area, including shots fired in an incident, and communicate information to dispatch, but they cannot detect or attribute gunshot activities specifically to an officer's service weapon. ¡ Technology advancements and independent testing, evaluation, and implementation research are needed to accelerate adoption. Technology developers are currently working through several technical hurdles and are leveraging insights from BWC to improve technology uptake. Some commercially available products have been evaluated for performance, but more studies are needed as technologies are further developed and released into the market.  

Criminal Justice Testing (and Evaluation Consortium, 2024. 15p.

By Camello, M. L., Houston-Kolnik, J. D., & Planty, M

This technology brief explores the use of chatbots within the criminal justice system. The goal of this brief is to orient the reader to chatbots, present foundational insights from real-world examples of chatbot use, highlight considerations for implementation, and discuss the future of chatbots in the criminal justice system  

  Key Takeaways ¡ There are numerous benefits to implementing chatbots, including: Ÿ improved efficiency for users accessing information, Ÿ enhanced community engagement by creating a 24/7 communications channel, Ÿ expanded access to justice through multilingual chatbot capabilities, Ÿ reduced costs by automating FAQ support traditionally done through live chat, and Ÿ reduced staff workloads. ¡ Chatbots carry inherent risks that decision-makers need to consider before implementation, including: Ÿ misinterpretation of user input leading to incorrect responses, Ÿ biased training data, and Ÿ vulnerability to hacking. ¡ Advancements in AI have enhanced and will continue to enhance chatbot capabilities and applications; however, despite these advancements, deploying AI-driven chatbots is not a “plug-and-play” opportunity for criminal justice applications

Research Triangle Park, NC:RTI International, 2021. 15p.  

By Redden, J., Aagaard,B., Taniguchi, T  

 This technology brief is the second in a four-part series that examines artificial intelligence (AI) applications in the criminal justice system. This brief highlights AI applications currently in use by law enforcement agencies, introduces frameworks for evaluating AI applications, and summarizes critical risks to consider when deploying AI systems. Additional briefs provide a high-level overview of AI within the criminal justice system and AI topics related to the criminal courts system and corrections.   

  Key Takeaways ¡ AI and advanced robotics in policing are not yet widespread; however, many law enforcement agencies are experimenting with these technologies. Opportunities to utilize AI applications in law enforcement will continue to increase as technologies evolve, including AI, 5G, and autonomous vehicles. ¡ This brief provides use cases, products, and vendor technologies to illustrate how some agencies have incorporated AI; the hope is for this information to inspire an ongoing dialogue between law enforcement leaders about how to improve policing. ¡ AI solutions hold promise to increase efficiency, promote data-driven practices, and expand capabilities for law enforcement agencies. The challenge will be for law enforcement agencies to identify use cases in which data quality and availability, technology maturity, and ethical constraints match their needs and their communities’ needs. ¡ Law enforcement agencies, communities, and the legal system need to have ongoing conversations about the tradeoffs between personal privacy and public safety/ security as AI enables more sophisticated surveillance and investigation

Criminal Justice Testing (and Evaluation Consortium, 2020. 10p.

By Smith, J., Camello, M., & Planty, M

  Generative artificial intelligence (AI) refers to AI1 used to create content, such as text, images, music, audio, and videos.2 Generative AI offers many potential benefits, enabling users to automate, augment, and accelerate a wide range of workflows, from simple administrative tasks like transcription and translation to more-complex functions such as investigation and decision support. In the criminal justice system, generative AI offers promising solutions to address human resource and budget challenges, allowing practitioners to focus on more-impactful work. Generative AI–integrated tools may enhance data analysis, improve detection and objective assessment of evidence, and streamline administrative processes. However, its integration, particularly in the criminal justice domain, raises some concerns, including potential biases, privacy issues, and the need for rigorous oversight to ensure effective implementation. It is unclear whether these tools can deliver on their promised efficiencies in practice, as evidenced by early research evaluating time savings of implementing AI-assisted report writing software.3 These concerns highlight the necessity for addressing bias and accuracy, maintaining strict data privacy and security protocols, and promoting transparency and accountability in AI-driven decisions and processes. This report is intended to help criminal justice decision-makers do the following: ¡ Understand what generative AI is and how it relates to the criminal justice system ¡ Identify how generative AI may be applied to tasks and jobs within the criminal justice system and the potential benefits, realities, and limitations ¡ Consider the technical, operational, and governance factors that may influence adoption and implementation ¡ Understand what makes up the generative AI technology stack and how models can be trained Key Takeaways • Generative AI represents an acceleration and advancement in technological innovation that already impacts the criminal justice system and will continue to do so—it is no longer a question of if or when, but how and to what extent. • Generative AI–powered software tools may offer many potential benefits, such as improving efficiency and augmenting capabilities across an extremely broad set of applications for criminal justice system stakeholders. Although these products hold promise, little empirical evidence currently supports or refutes promised benefits from these products. • Generative AI models can be deployed in various forms, including cloud-based models that centralize data processing and federated models that enable decentralized training across multiple locations, preserving data privacy and enhancing security for sensitive criminal justice applications. • Decision-makers should be aware of the substantial technical, operational, and governance risks associated with generative AI– powered software tools prior to implementation. • Responsible use of generative AI requires addressing bias and accuracy concerns, maintaining strict data privacy and security protocols, adhering to ethics and legal standards, and promoting transparency and accountability in AI-driven decisions and processes. • Generative AI technology is evolving faster than the legal or policy environment for AI—the criminal justice community must be proactive and must implement robust internal training and policy frameworks rather than relying solely on external legal or regulatory guidance.  

Research Triangle Park, NC: RTI International, 2025 28p.