This chapter addresses algorithmic bias and fairness in applied AI systems across high-impact domains such as criminal justice, healthcare, employment, and finance. It examines how biases—stemming from unbalanced data, historical inequalities, and flawed design choices—can lead to discriminatory outcomes that affect individuals’ access to rights and services. Four core types of bias are outlined: sampling, labeling, measurement, and historical bias. The chapter covers technical and governance-based strategies to mitigate these risks, including fairness-aware preprocessing, adversarial debiasing, and the use of fairness metrics such as demographic parity and equalized odds. It emphasizes the importance of transparency and accountability through the use of interpretability techniques like LIME and SHAP, which reveal how input features influence model decisions.