The use of artificial intelligence (AI) in the legal system has the capacity to improve efficiency, consistency, and accessibility. There are, however, issues around bias and fairness in AI-led legal decision-making that pose profound ethical and legal issues. The research discusses the way biases in training data, algorithmic structure, and human intervention affect the fairness of AI-based legal systems. It investigates issues such as discriminatory outcomes, non-transparency, and accountability in automated decision-making. Through case studies of AI use in legal settings, such as predictive policing, sentencing suggestions, and automated contract review, the study points to cases where bias has resulted in unfair legal outcomes. It also examines regulatory protocols and ethical standards intended to reduce such bias while maintaining adherence to core legal principles such as due process and non-discrimination. The study also assesses the function of human oversight in reducing AI bias, considering the efficacy of interventions like algorithmic auditing, explainability mandates, and fairness-promoting design methods. It addresses the conflict between the efficiency of AI-based systems and the necessity of judicial discretion, highlighting the significance of preserving ethical protection in automated legal decision-making. Ultimately, this study is a contribution to current discussions of responsible AI design and legal ethics, offering policy suggestions to ensure fairness, accountability, and trust in AI-powered legal systems. It emphasizes the need for cross-disciplinary work among technologists, legal experts, and policymakers to develop AI frameworks that protect justice, equality, and fundamental rights in the changing context of legal automation.