Algorithmic bias in AI, particularly across emerging economies like India, presents complex challenges that go beyond technical issues. Too often, AI unintentionally deepens social divides—whether by caste, gender, class, or geography—complicating efforts toward fairness. This chapter proposes a model of Compassionate Governance with three layers—Prevention, Detection, and Healing—that leaders can use to guide fair and just AI use. This approach asks leaders to reflect on real human impact, involve diverse voices in design, and react quickly when things go awry. Bringing in current studies, field cases, and classic leadership thinking, this chapter highlights why being ethical and compassionate in AI adoption is essential for trust—and for preventing harm.