Biased AI algorithms predicting risk scores of recidivism negatively impact specific demographics since results are used by probation departments. For instance, individuals with higher risk scores have been found to be more likely to be detained before their trials.
Our goal is to create an AI model that classifies the risk of recidivism with minimal bias. We specifically aim to obtain a model satisfying counterfactual fairness for the scope of this hackathon. Moving forward, we hope to build a model mitigating all measures of biases to be utilized by law enforcement.
Log in or sign up for Devpost to join the conversation.