Abstract

As AI systems increasingly influence critical decisions in sectors like healthcare, recruitment, finance, and criminal justice, concerns around bias and fairness have intensified. Although often viewed as neutral, AI can unintentionally replicate or amplify social inequalities due to biased training data, narrow performance metrics, or limited contextual design-disproportionately impacting marginalized communities. This study underscores that AI ethics is fundamentally a human issue affecting dignity, trust, opportunity, and justice. By examining fairness frameworks and real-world cases of algorithmic bias in hiring, healthcare, and predictive policing, it reveals how well-meaning systems can produce harmful outcomes. The paper also critiques existing mitigation strategies for addressing only surface-level issues while overlooking deeper structural inequities. It advocates for a human-centered AI approach that emphasizes transparency, inclusivity, stakeholder participation, continuous auditing, explainability, and accountability. Ultimately, the study calls for aligning AI development not just with technical capability, but with core human values such as fairness, responsibility, and equity to ensure just and sustainable societal impact.

Keywords:AI ethics; Algorithmic bias; Fairness in AI; Ethical decision-making; Human-centered AI; Responsible AI; Machine learning fairness; Data bias; Algorithmic accountability; Transparency in AI; Social impact of AI; Inclusive AI systems; Justice in AI; AI governance; Moral responsibility in AI