Mauro Zamboni (Stockholm University – Faculty of Law) has posted Artificial Intelligence in Legislative Governance: A Legal-Institutional Framework for Allocating AI on SSRN. Here is the abstract:
Artificial intelligence is becoming increasingly embedded in the processes through which legislation is prepared, adopted, implemented, evaluated, and revised. The central legal question is therefore no longer whether AI should be used in legislation, but what role it may legitimately perform within legislative governance. This article develops a legal-institutional framework for answering that question by focusing on the institutional function that AI is expected to support. It distinguishes between epistemic functions, through which legislative institutions acquire, organise, analyse, compare, and communicate knowledge, and legislative judgment, through which democratically authorised institutions determine the future direction of the law, balance competing public interests, publicly justify normative choices, and assume responsibility for them. AI can substantially strengthen the former, while the latter should remain with institutions exercising legislative authority. The distinction is further reinforced by the future-constitutive character of legislation and the past-dependent epistemic capacity of AI. Applying the framework across the adoption, implementation, and evaluation phases of legislative governance, the article shows that AI’s legitimate role varies according to the institutional function performed at each stage. Participatory democracy provides a further criterion for determining how otherwise legitimate AI assistance should be organised. The article ultimately argues that AI should strengthen the knowledge upon which legislative decisions are made without becoming the source of those decisions or of their democratic legitimacy.
Recommended!
To receive new posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.
Lawrence Solum