This article examines how transparency can function as a foundational safeguard in algorithmic governance within the European legal space. The Council of Europe’s Framework Convention provides a broad, rights-oriented vision of transparency grounded in intelligibility, oversight and participation. In contrast, the EU AI Act adopts a predominantly technical, market-driven approach, offering complementary but more limited transparency obligations. The paper argues that national administrative law plays a crucial role in operationalising these frameworks by articulating transparency as a multi-layered principle – combining access to information, reason-giving, participation and review – to address the structural opacity of AI systems. Through this administrative infrastructure, algorithmic
decision-making can be reconciled with democratic accountability and the rule of law.