Bias in AI happens when AI systems make unfair or discriminatory decisions. It often originates from biased training data, reflecting societal prejudices or historical inequalities. Fair data is crucial because AI algorithms learn from what they're fed; biased data leads to biased outcomes. Recognizing and mitigating bias requires careful data governance, diverse teams, and continuous monitoring. Addressing bias is essential for building trustworthy AI, ensuring ethical compliance, and achieving equitable societal impact.