•  
  •  
 

Abstract

The dominant academic literature on artificial intelligence and fundamental rights has been organised around a structuring premise: that AI systems pose a systemic threat to constitutional values, and that rights protection is a project of constraining and resisting automated decision-making. This article accepts and supports the empirical foundation of that literature. It does not accept that the framework is complete. This article proposes the concept of AI-augmented constitutionalism, defined as the deployment of artificial intelligence systems to strengthen the detection, enforcement, and access dimensions of fundamental rights, subject to a rule of law safeguards architecture. The concept inverts the dominant analytical frame: instead of asking what AI does to rights, it asks what AI might do for them. The claim is explicitly conditional. The properties that make AI constitutionally dangerous are directional, not intrinsically antagonistic to rights. The article tests this framework against three fundamental rights, namely: equality and non-discrimination, freedom of speech, and the right to education, through three thematic dimensions: detection, enforcement, and access. The principal comparison is between the United States and the European Union, but the article argues that a framework aspiring to general theoretical significance cannot rest on a transatlantic comparison alone. Global South jurisdictions contribute three analytically distinct lessons: constitutional architectures with stronger positive rights foundations that create legal space for AI-augmented enforcement the US order cannot easily accommodate; empirical tests of the framework at a scale and under conditions of institutional constraint that sharpen its conditional structure; and a limit condition, the systematic weakness of AI tools in non-English linguistic environments, that the transatlantic literature has not yet confronted. The normative anchor is the substantive rule of law, organised around four pillars: transparency, accountability, contestability, and rights-compatibility by design. The article acknowledges four genuine limits of the framework and concludes that AI-augmented constitutionalism is a political and institutional achievement rather than a technical given, one that requires active construction, not passive deployment.

Digital Object Identifier (DOI)

10.55496/SPUE2081

Download

Share

COinS