Artificial Intelligence Adoption and Governance in Telecommunications: A Literature Review of Drivers, Ethical Risks, and Governance Models in Emerging Markets

Abstract

Telecommunications operators were among the earliest adopters of artificial intelligence (AI) in network operations, and the sector now offers one of the richer empirical literatures on what drives AI adoption and how its governance is beginning to be formalised. This review synthesises twenty peer-reviewed studies to examine four questions: what factors drive AI adoption in telecommunications firms, particularly in emerging markets; where AI is applied within network operations, including the specific and commercially consequential case of customer churn prediction; what risks accompany AI's expansion into 5G network management specifically; and what governance and ethical risk frameworks the literature proposes for managing AI's expansion into safety, privacy, and competition-sensitive functions. Using the Technology–Organisation–Environment (TOE) framework that recurs across studies from China, Indonesia, Sri Lanka, and India, the review finds that managerial support and organisational readiness are consistently stronger predictors of adoption than external market pressure. AI is most mature in predictive maintenance, churn prediction, and customer-facing analytics, and least mature in autonomous network management, where governance frameworks are still emerging, notably an ambidextrous COBIT-based model tested in an Indonesian operator, a set of 5G-specific risk and governance frameworks, and a series of ethical leadership frameworks proposed for AI-enabled service industries generally. The review closes by considering what these findings imply for AI governance in Somalia's telecommunications sector, a sector with unusually mature mobile money regulation but, on current evidence, no documented AI-specific governance framework of its own.