Artificial Intelligence in Parliamentary Activity: Institutional Governance, Transparency and Accountability

Abstract

The article examines the legal and institutional conditions under which artificial intelligence may be integrated into the activity of a national parliament without eroding the constitutional nature of legislative decision-making. Parliament is analysed in its dual capacity as a regulator of artificial intelligence and as a user of it, and the principle of reflexive consistency is substantiated, according to which the standards that the legislature imposes on other actors must bind its own use of algorithmic tools. Drawing on the resolution and Guidelines of the Inter-Parliamentary Union, the Council of Europe Framework Convention and the European Union Artificial Intelligence Act, three governance models are distinguished in comparative practice: the administrative-policy model of the United States Congress, the in-house development model of the Brazilian Chamber of Deputies and the low-risk-first model of the European Parliament and several European legislatures. The national framework of Uzbekistan, formed by the Law of 21 January 2026 on relations arising from the use of artificial intelligence and the programme of the Council of the Legislative Chamber of 13 April 2026, is assessed against these benchmarks, and proposals are formulated concerning a joint act of the chambers, the advisory status of algorithmic expertise, a public register of parliamentary artificial intelligence systems and the parity of deputies' access to such tools.