OECD Policy Paper: Intellectual Property Issues in Artificial Intelligence Trained on Scraped Data

Recent advancements in artificial intelligence (AI), particularly the emergence of generative AI, have sparked discussions about the evolving intellectual property (IP) landscape. Some data collection methods create questions about intellectual property protection and other relevant rights as the need for AI training data increases. 

The Organisation for Economic Co-operation and Development (OECD) recently published a new policy paper providing an overview of key issues at the intersection of AI and some IP rights. The paper seeks to enhance understanding of data scraping, a fundamental method for acquiring AI training data used in developing large language models. It examines data scraping techniques, maps key stakeholders, and explores global legal and regulatory responses. Lastly, it presents initial considerations and potential policy approaches to assist policymakers in balancing AI’s innovation potential with the protection of IP and other rights.