Abstracting Injustice: An Analysis of the Use of Artificial Intelligence in Criminal Justice
Vincent Huynh-Watkins
Committee: Stephen Fickas
Honors Bachelors Thesis(Jun 2021)
Keywords: AI Ethics, Machine Learning, Criminal Justice,

As artificial intelligence (“AI”) has become commonplace in many aspects of life and society–often seen as a faster, more accurate, and less labor-intensive alternative to human cognition –the use of AI in criminal justice systems has been a naturally occurring phenomenon. There are many potential applications of AI in criminal justice that may seem sensible, with the touted possibility to provide fairer outcomes and increased safety for society. For instance, AI-based facial recognition may help investigators and prosecutors solve previously unsolvable cases. AI may also help law enforcement agencies predict criminal activity prior to its occurrence (known as predictive policing), which may help in resource allocation and targeting areas for increased policing, theoretically leading to reduced rates of crime. Further, some researchers claim that AI algorithms can provide a more objective and complete analysis of the recidivism risk posed by convicted criminals, therefore providing a better basis for sentencing to prevent repeat offenses and free those who are not repeat threats. These are just some of the potential uses of AI in the justice system, however they are seen as particularly promising applications of technology which have the potential to make society safer and fairer and are beginning to flood the world of criminal justice. In this paper, the implications of AI use in criminal justice are explored and reviewed followed by analysis on how these emerging issues ought to be addressed. Ultimately this is an intersectional and interdisciplinary analysis of power, computer science, the criminal justice system, and where AI fits into this puzzle; what its role is and how it has shaped and will continue to shape wide-ranging outcomes in a variety of applications.