A legal study of transparency and fairness in algorithm-driven legal decision support systems

Nowadays, more and more legal decisions are being made with the help of machine learning models and algorithmic drives. Avoiding or reducing human bias in the decision-making process is extremely important for the fairness and transparency of the algorithms in legal decision support systems, and also profoundly affects the fairness and transparency of legal support systems. The article develops a system for supporting legal decisions based on the decision tree. By improving the traditional decision tree algorithm, the decision tree can be used to mine and classify unknown legal text data samples. Finally, the transparency and fairness of the legal decision support system proposed in this paper are legally researched, and legal decision accuracy is measured. Through experimental research, this paper concludes that the average accuracy of decision-making in commercial and copyright cases is 89% and 91%, and overall, the legal decision support system based on the decision tree proposed in this paper has good decision-making accuracy. © 2024 Shengji Rong, published by Sciendo.