FROM COMMONS TO CODE: HOW AGENTIC ARTIFICIAL INTELLIGENCE IS FORCING A POLYCENTRIC REDESIGN OF IP LAW

Abstract

The digital economy, expected to reach a worth of US $ 24 trillion by late 2025, faces an Incentive-Access Paradox of the kind that is an unavoidable juxtaposition of the difficulty intellectual property frameworks have in coping with the rise of Agentic artificial intelligence and the decentralised architectures that support it. Established anthropocentric doctrines of authorship and inventorship conflict with the products of autonomous machines, thereby engendering concerns about a “Tragedy of the Anti-commons” in which fragmented rights and patent thickets hamper the very innovation they seek to protect. Regulatory efforts around the world are increasingly oriented toward systemic risk management and hybrid licensing regimes.

The European Union’s AI Act, which comes into partial effect in August 2025, includes mandates for transparency and copyright compliance for providers of General-Purpose Artificial Intelligence (GPAI). In parallel, the 2025 Directive of the Indian Ministry of Industry and Commerce (DPIIT) expert committee has favored a “mandatory blanket licence” accompanied by statutory remuneration; it has thus effectively sought to balance the disparity in bargaining power between the developer/creator. The judicial frontier has been further identified by US precedent set in 2025, specifically Bartz v. Anthropic and Kadrey v. Meta. These decisions characterised training of Large Language Models (LLMs) as “spectacularly transformative” and thus fell within the ambit of fair use. However, the rulings set the transformative process of training apart from unlawful data acquisition, which keeps liability for using pirated sources alive. In an attempt to fill the gap in today’s law, sui generis mechanisms, e.g., Ukraine’s 25-year-long-term economic protection regime for non-original computer-generated works, are emerging as viable alternatives to mainstream copyright law.

Developing a resilient digital ecology depends on a polycentric governance structure that addresses the Productivity J-Curve and prevents the enclosure of the digital commons, which is essential for providing legal certainty and supporting the US $6.7 trillion infrastructure investment needed by 2030