TRENDS Research & Advisory has released the second issue of its “AI TRENDS” series, titled Algorithmic Governance and the Legitimacy of AI-Supported Decisions in the Public Sector. The issue addresses a central challenge: the widening gap between governments’ continued adoption of artificial intelligence (AI) and the slow development of regulatory and institutional frameworks that ensure legitimacy.
The study argues that the legitimacy of algorithmic decisions in the public sector is no longer solely a legal issue, but has become a multidimensional concept encompassing law, technology, administration, institutions, and societal trust. Drawing on a comparative analytical approach that reviews the literature, international experiences, and frameworks, the study finds that challenges to legitimacy stem not only from technical shortcomings but also from institutional governance. Foremost among these are the “black box” problem, limited transparency and explainability, algorithmic bias, ambiguity surrounding accountability and legal responsibility, and the risk of privacy violations.
Specialized Legislation
The study proposes a comprehensive Arab governance framework based on enacting specialized legislation, establishing oversight structures, making algorithmic impact assessments mandatory before deploying high-risk systems, and ensuring human oversight and independent auditing, to balance innovation with rights and the rule of law.
The study explains that legitimacy depends on the harmonious integration of five interconnected dimensions: legal, procedural, technical, institutional, and societal. Its principal scholarly contribution is to formulate the concept of “algorithmic legitimacy” as an extension of administrative legitimacy and to present an original framework, the “Algorithmic Legitimacy Pyramid,” for systematically assessing decisions and managing their risks.
“Algorithmic governance” is a central concept in the literature on public administration and digital governance. It refers to the frameworks, rules, and institutions that regulate the design, use, and oversight of algorithmic systems within the public sector. As government decision-making increasingly relies on artificial intelligence, algorithmic governance has become essential to balancing two competing objectives: harnessing technology’s potential and safeguarding the fundamental values of the rule of law.
Legal and Institutional Challenges
The study stresses that the public sector is undergoing a fundamental transformation in decision-making as AI technologies and algorithms rapidly expand in managing public services, allocating resources, assessing risk, and supporting policy. Despite the considerable potential of these technologies to enhance operational efficiency and accelerate procedures, they have also introduced new legal, institutional, and ethical challenges, raising questions about the legitimacy of government decisions that directly affect individuals’ rights, freedoms, and fundamental interests.
It notes that this transformation raises a range of legal, institutional, and ethical questions concerning the nature of the authority exercised by algorithms, the limits of its legitimacy, and the mechanisms for holding it accountable. Government decisions derive their legitimacy not only from technical accuracy, but also from their appropriateness and conformity with the principles of legality, justice, transparency, accountability, and public trust.
Legitimacy of Algorithmic Decisions
The study concludes that algorithms have become influential actors in producing administrative decisions. The question is therefore no longer limited to how efficient these systems are, but extends to the legitimacy of the decisions they produce. International experiences in recent years have revealed the challenges associated with government use of AI. Some algorithmic applications have produced discriminatory or unfair outcomes and raised issues related to privacy, accountability, and individual rights. These experiences have shown that technical success does not necessarily guarantee societal acceptance or institutional legitimacy, and that failing to manage the risks of artificial intelligence may lead to wide-ranging political and legal crises.
