Digital discretion and public administration in Africa: Implications for the use of artificial intelligence

African public administrations are in an especially difficult position with respect to the adoption of artificial intelligence (AI), both how they harness its benefits but also managing potential harms. Through a narrative review existing research, this paper synthesises findings from previous digital government implementations on continent, and considers implications for AI use. The analysis is centred way officials bypassed or empowered work emerging technologies, influence service values their adoption. Four concerns emerge as relevant: Integrity recommendations provided by decision-support systems, including influenced local organisational practices reliability underlying infrastructures; Inclusive decision-making that balances (assumed) objectivity data-driven algorithms different stakeholder groups; Exception accountability platforms funded, developed, implemented used; expectation Complete understanding people events through integration traditionally dispersed data sources policy actors seek mitigate risks associated aspiration.