Abstract
The rapid expansion of AI in higher education creates major opportunities, yet under-resourced institutions face heightened risks of inequitable and unethical adoption. A comprehensive meta review of studies conducted in higher education settings revealed that 40.9% of the studies reported ethical gaps, 29% cited lack of ethics as a primary barrier, and 65% indicated weak foundations without frameworks; misuse persisted, with 17% of academic papers containing ≥20% AI-generated content and 5% exceeding 80% (Bond et al., 2024; Bennett et al., 2025; Rafiq & Qurat-ul-Ain, 2025). Readiness gaps are substantial: 92% of students use AI, but only 36% receive training, about 10% of institutions provide guidance, 78.5% of faculty need training, and 85.8% report their institutions do not promote ethical AI (Alnsour et al., 2025; Mazaheriyan & Nourbakhsh, 2025; UNESCO, 2023). Grounded in major developmental, sociocultural, engagement, and equity theories, this dissertation shows structured AI ethics instruction is a high-impact practice linked to stronger engagement, ethical reasoning, and learning outcomes, with 80% of faculty supporting ethics instruction, 51% of studies reporting positive effects, and 32.3% improved outcomes (Alnsour et al., 2025; Bond et al., 2024). The project therefore introduces an open-access, cross-platform AI Literacy and Ethics Toolkit to strengthen institutional capacity, support responsible AI integration, and promote equitable student success.