Artificial intelligence (AI) is being absorbed into the legal and administrative institutions of the Global South faster than the doctrinal frameworks governing those institutions are being revised. This review examines that gap comparatively, using India and Nigeria as paired cases. It synthesises eight recent studies published in the British Journal of Contemporary Research alongside the wider governance literature, organising the evidence around four themes: conceptual boundary-setting for legal purposes; AI in adjudication and procedural justice; AI in executive and administrative decision-making; and the infrastructural, labour-market and cybersecurity preconditions on which any regulatory design depends. Three findings emerge. First, both jurisdictions have converged on principle-based, non-statutory governance instruments India's AI Governance Guidelines and Nigeria's National AI Strategy that articulate transparency and accountability commitments without creating enforceable individual remedies. Second, the binding constraint in both cases is institutional rather than doctrinal: uneven digitisation, fragmented data, thin technical audit capacity and low digital literacy limit what any legal standard can practically achieve. Third, the two states are diverging in emphasis, with Indian scholarship concentrating on constitutional and procedural review of algorithmic power and Nigerian scholarship concentrating on infrastructure security and labour transition. The review argues that rights-based safeguards and capacity-building are complementary rather than sequential, and identifies the absence of empirical evidence on deployed public-sector systems in both jurisdictions as the most consequential research gap.