This letter features reporting from “How AI-Powered Tech Landed Man in Jail With Scant Evidence” by Garance Burke, Martha Mendoza, Juliet Linderman, and Michael Tarm, a Pulitzer Center reporting project
Dear Commissioner Joe Martinez,
Wrongful imprisonment has threatened the lives and well-being of many Americans. There have been over 2,500 documented exonerations since 1989, and these innocent people have suffered immense mental trauma due to these false convictions. The article “How AI-Powered Tech Landed Man in Jail with Scant Evidence,” published by the Pulitzer Center, describes the effect that ShotSpotter, a gunshot detection tool, has had on wrongful imprisonment. The authors explained the case of Michael Williams, who was jailed based on evidence from a ShotSpotter sensor, and then had his case dismissed due to the insufficient nature of the ShotSpotter evidence. ShotSpotter sensors have missed shots in close proximity and misclassified other loud sounds as gunshots, having a low efficacy rate. Crucial data, records, and the algorithm behind the ShotSpotter system have been shielded from the public, leaving the validity and reliability of the technology in question. Additionally, sound classification and location have been edited by ShotSpotter employees, with several known cases described in the article.
The use of ShotSpotter is part of a global issue regarding the use of AI implemented in government and workplaces. AI systems have been used in many countries, creating the framework for policies in the justice system, selecting job applicants for consideration, and enforcing laws. Their widespread use, including systems in China and the U.S., has led many to question the efficacy of the algorithms behind these systems. Researchers at an MIT lab found that these algorithms have low efficacy rates for both women and people of color. The weaknesses of these algorithms often entrench the systemic oppression of women and people of color in the workplace and justice system, and lead to higher rates of wrongful imprisonment, just as seen in Michael Willams’s case. The implementation of ShotSpotter furthers the effects of these algorithms because the sensors have been placed in communities deemed to be more “at-risk,” which tend to contain large populations of African-Americans and people of color. This places these communities at greater risk of wrongful imprisonment and increased scrutiny from false gunshot detections, given ShotSpotter’s high rate of error.
These issues have recently extended into our community. In 2016, the same year you were elected, the Board of County Commissioners unanimously voted to reinstate the use of ShotSpotter in Miami-Dade. ShotSpotter was previously employed by the Miami-Dade Police Department in 2012, before being discontinued as it wasn’t considered effective. While the company argues that the algorithm and machine learning model is constantly being fine-tuned, experts have warned that their attempts at improvements may lead to increased flaws and inaccuracy, building “inherent uncertainty into their system.” According to the Miami New Times, the police department did not disclose the previous findings that the county has had in employing ShotSpotter to the Board that was voting on the use of the system. Commissioner Martinez, due to your 17-year commitment with the police force, you must understand the importance of keeping our officers and community safe. It is with these principles in mind that I strongly encourage you and your colleagues to re-evaluate the allocation of up to $5.6 million to the use of ShotSpotter in Miami-Dade.
The money used for ShotSpotter can be reallocated towards social workers and improving the community environment in at-risk areas. There is a strong correlation between poverty and crime rates, and improving living conditions and education can decrease the amount of crime seen in at-risk youth. This would not only improve the community, but protect the police force by reducing situations in which they are put in danger. This is a solution that benefits both the community and improves the safety of the officers, and you should consider these reallocations due to the lack of efficacy seen in the ShotSpotter system.
Thank you for your consideration,
Kavita Doobay
Kavita Doobay is a Junior at TERRA Environmental Research Institute. She is passionate about volunteering and getting involved in her community, and she has done this by starting a robotics program at Frank C Martin K-8 Center, teaching a children's Hindu Philosophy class at the Miami Lakshmi Narayan Mandir Bal-Vihar, and advocating for autism acceptance with Luv Michael. Her hobbies include going on long walks with her family, reading, and baking. She'd like to thank her friends, family, and teachers who support her through her academic and personal endeavors.
Read more winning entries from the 2021 Local Letters for Global Change contest!