ShotSpotter gunshot detection 

ShotSpotter is an acoustic gunshot detection system manufactured by US-based company SoundThinking (named ShotSpotter until April 2023). Consisting of microphones, sensors, algorithms, and human reviewers, the system alerts police to potential gunfire.

SoundThinking claims 97 percent accuracy and 0.5 percent false positive rates across its ShotSpotter customers, and has been used by multiple municipal authorities, police departments and school districts across the US since 1997. 

It has also been used as evidence in legal trials.

Risks and harms 🛑

ShotSpotter has proved controversial, with concerns expressed about its accuracy, reliability and effectiveness, and the risks it is seen to pose to human rights and civil liberties and community relations through surveillance and over-policing.

SoundThinking has also been accused of wasting police time and resources, misleading marketing, and displaying poor transparency and inadequate accountability.

Transparency and openness 🙈

Despite being mostly used for public safety, ShotSpotter has been criticised for poor transparency across multiple dimensions, leading to questions about its effectiveness and to accusations of misleading marketing and poor accountability.

• Proprietary algorithm. Despite claims that the ShotSpotter system has 'serious flaws' regarding its accuracy and false positives, SoundThinking refuses to disclose its exact methods for detecting and locating gunshots, claiming it is a trade secret. As ShotSpotter updates its algorithm, it is unclear how these changes might affect its performance, accuracy or reliability over time. Furthermore, in court cases where ShotSpotter evidence has been used, defense attorneys have argued that they cannot properly challenge the evidence without understanding how it was generated. 

• Data access. Researchers often face challenges in accessing comprehensive data from police departments and ShotSpotter, limiting the scope and depth of potential studies. Accordingly, independent research - including peer-reviewed academic studies - on ShotSpotter's impact on solving crimes, preventing gun violence or the system's impact on community-police relations or residents' perceptions of safety is limited, making it difficult for policymakers and the public to objectively assess whether ShotSpotter is an effective tool for reducing gun violence and improving public safety. It also complicates decisions about allocating resources to this technology versus other crime prevention strategies. Rather, much of the available data on ShotSpotter's effectiveness comes from studies funded or conducted by the company itself, raising questions abput conflicts of interest and potentially introducing bias. And most of these studies focus on short-term impacts, while the long-term effects of using the technology are less clear.

• Data retention. Questions exist about how long SoundThinking keeps ShotSpotter audio data, who can access it and under what circumstances - a lack of transparency that raises privacy concerns as the system may capture conversations or other sounds beyond gunshots. It also raises concerns about how gunshot and personal data could be used for other types of surveillance or investigations.

• Human intervention. While ShotSpotter claims that human reviewers validate each alert before sending it to law enforcement, the exact process and criteria used for this review are not publicly detailed, nor are details about the background and training of its human reviewers, their ability to override the system and error rates, and the accountability of their decisions.

• Contract terms. Details of agreements between ShotSpotter and cities are often bound by non-disclosure agreements (NDAs), preventing full public disclosure of the terms, cost structure, performance guarantees, data ownership, liability, audit rights, data sharing and other elements, making it difficult for authorities and the general public to assess the true cost and value of the technology, and hindering public oversight and accountability.

• Bias and privacy abuse. Concerns exist about how the lack of information and public engagement on the placement of sensors may affect different communities - including the criteria used to place them, the concentration of their use in areas with higher reported gun violence, which frequently correlate with low-income and minority neighborhoods and can lead to racial profiling, the over-policing of these communities, and the erosion of trust in law enforcement. Concentrating sensors in certain areas may also skew gun violence statistics, po