My original research plan this year focuses on mimicking crime forecasting & risk assessment algorithms currently used in the legal justice system.
My original research this year has the following steps:
- Finding a dataset that a light question can be written based off of (e.g. Which is your favorite superhero)
- Coming up with the exact question
- Writing a survey with questions based on the dataset
- Writing and training an algorithm based on the dataset’s breakdown
- Giving the survey to the student body at Renaissance and then running it through the algorithm
- Analyzing the results my algorithm got versus what the user said
My hope is that this will draw parallels between how flawed algorithms can be and how flawed they are in the context of race, gender, and other marginalized factors, especially in the legal justice system.