Artificial Intelligence and Predictive Justice

Défense de mémoire Otri Dania

Rejoindre la défense

 

Much has been written recently about artificial intelligence (AI) in the field of justice. However, is it safe to let AI decide what is best for mankind? I would think not. AI can ease and pave the way to make a decisive and righteous decisions. Therefore, “AI needs social scientists to ensure AI alignment algorithms succeed, when actual humans are involved.” (Geoffrey Irving & Amanda Askell). This thesis explains the function of the artificial neural networks and how it learns to make better decisions. Through a database processing and deep learning, ANN can determine the decision by comparing the case to similar cases using a dataset. AI is deprived from morals and emotions, thus it’s able to give an objective decision. On the other hand, this might cause an ethical dilemma, therefore, there must be a human judge to give the final decision. In this study, AI judge would be applied on a collective settlements of debts case. The discussion aims to be nuanced but also understandable to those without a technical background. To that end, First, I’m going to discuss Justice in general, then turn to AI and how it’s being used in the practice of justice and government officials, who administer the law. I’ve made two software to demonstrate the roll of AI in the field of law and justice. First software is Decision Making Assistant which calculates and gives results the judge needs to make the right decision. Second software depends on deep neural network that does all the calculations and present final decision without any intervention by a judge. Key motivation in writing this thesis is to provide a realistic, demystified view of AI that is rooted in the actual capabilities of the technology.

 

Keywords:  artificial intelligence (AI), deep neural networks (DNN), predictive justice, decision making assistant and judge.

Contact : Isabelle Daelman - 5252 - isabelle.daelman@unamur.be
Télecharger : vCal