Rethinking Parliamentary Autonomy in the Age of AI: A Study of Digital Constitutional Theory

Abstract

The principle of parliamentary autonomy is a cornerstone of representative democracy. Constitutions require parliaments to make political decisions independently of any other power. Traditionally, constitutional scholarship has conceptualized parliamentary autonomy in relation to the risk of undue influence from other branches of the state. However, within the emerging field of digital constitutionalism, scholars have noted that constitutional challenges are increasingly shifting in nature. Against this background, the working paper examines how the use of AI in parliaments may affect their autonomy. It argues that drawing a parallel between AI systems and human experts can help constitutional theorists see that many challenges commonly attributed to AI-such as opacity and bias-are not entirely new, but rather inherent to broader informational practices. These characteristics are structural features of knowledge production itself and are thus also reflected in broader institutional dynamics that shape parliamentary functioning, such as bureaucratization and the scientification of politics. If constitutions allow, and at times even require, a degree of expert input in democratic governance, then similar reasoning should inform our understanding of AI. From this perspective, some critiques currently directed at AI may be more appropriately extended to human experts, whose epistemic authority within democratic processes is often accepted with far less scrutiny. The analogy fails, however, in those limited cases where AI-generated knowledge cannot be verified or meaningfully contested by human experts, situations in which a distinct constitutional risk arises. At the same time, AI may offer opportunities to enhance parliamentary autonomy-an aspect that remains largely unexplored in legal scholarship. In particular, AI could help reduce the information asymmetry that parliaments face, both in relation to other branches of the state (notably the executive) and to private actors (such as epistemic communities and the media). In this regard, AI technologies might enable more direct access to information relevant to public decision-making, ultimately strengthening parliamentary assemblies' independence when certain institutional safeguards are in place.