This report presents the case of Criminal Sentencing AI technologies as a research project, outlining the background and rationale to suggest a research plan. There has been a rise in the use of criminal sentencing AI technologies within the court system, with eight states incorporating recidivism scores into their sentencing processes. These states are Arizona, Colorado, Delaware, Florida, Kentucky, Louisiana, Oklahoma, Virginia, Washington, and Wisconsin. While AI holds promise, it can also inherit biases from its creators. When determining an individual's sentence, it is crucial to consider both factual data and the context of the specific incident. Relying solely on criminal sentencing technologies eliminates due process and further removes the "human" element from an already minimal judicial system. The question arises whether these tools disproportionately harm individuals and whether they are genuinely helpful or successful in predicting recidivism rates. In this research project, I aim to evaluate the effectiveness of the Compas tool while also examining the need for a "human" presence in the judicial system, particularly concerning the sentencing of individuals.
