Model bias in Artificial Intelligence (AI) systems poses significant challenges to reliability, fairness, and ethical alignment. Biases can arise from issues in training data, algorithm design, or deployment context, leading to skewed or discriminatory results. The underrepresentation of certain demographic groups in the training data, such as individuals aged 65 years and older or widowed individuals, raises concerns about inclusivity and equity. Addressing these biases requires comprehensive governance frameworks that span the entire AI lifecycle. Recent legal cases and disputes have highlighted the pervasive issue of bias in AI/ML models, emphasizing the need for fair and unbiased systems. Strategies for detecting and mitigating bias include data analysis and preprocessing, fairness metrics, model interpretability techniques such as SHAP and LIME, fairness-aware regularization, post hoc analysis, and continuous bias detection. These approaches aim to identify and address biases across the AI lifecycle, ensuring compliance with legal and regulatory standards while fostering trust among stakeholders. As AI technologies continue to shape industries and societies, prioritizing fairness and equity is essential to unlock their full potential, while safeguarding their ethical integrity and societal impact.