Generative AI and Academic Integrity in Online Distance Learning: The AI-Aware Assessment Policy Index for the Global South

Background: Generative artificial intelligence (AI) is transforming assessment and academic-integrity practice, yet most online and distance-learning (ODL) universities in the Global South must navigate this disruption with limited resources and acute equity concerns. Purpose: Guided by sociotechnical-systems theory, neo-institutional isomorphism, and value-sensitive design, the study synthesises the emerging landscape of AI-responsive assessment policy and interrogates its implications for integrity, inclusiveness, and epistemic justice. Design/Methodology: A secondary-data meta-synthesis integrated 70+ peer-reviewed articles, global policy reports, and 50 institutional documents released after 2022. The team created the AI-aware Assessment Policy Index (AAPI), a six-dimension composite that quantifies the orientation of institutional responses along a redesign–surveillance and transparency–opacity continuum. Forty ODL institutions across Africa, Asia–Pacific, and Latin America were scored and correlated with publicly reported indicators such as AI-detector flag rates and reliance on remote proctoring. Findings: Policies clustered into two archetypes. “Redesign-focused” frameworks emphasised authentic assessment, mandatory disclosure of AI assistance, equity safeguards, and AI-literacy programmes; “surveillance-heavy” regimes relied on automated detection and high-stakes proctoring. Higher AAPI scores were moderately associated with lower AI-plagiarism flag rates (Spearman ρ = –0.42, p < .01) and reduced proctoring dependence, without evidence of increased misconduct. Nonetheless, pervasive bias in AI-writing detectors and racialised facial-recognition errors expose systemic risks that disproportionately burden multilingual and low-bandwidth learners. Practical Implications: The findings support a shift from reactionary policing to design-led governance. The article offers a nine-step policy blueprint—spanning assessment redesign, procedural safeguards, and capacity-building—that ODL leaders can adapt to local contexts. Originality/Value: By operationalising an empirically validated policy index and embedding post-colonial critique, the study provides the first comparative evidence that integrity in the AI era is best safeguarded by human-centred, transparent, and context-aware assessment design rather than escalating surveillance.