Intellectual property issues in artificial intelligence trained on scraped data

As the demand for AI training data surges, certain data collection methods raise concerns about the protection of intellectual property (IP) and other rights. This overview of key issues at the intersection of AI and some IP rights facilitates a greater understanding of data scraping, a primary method for obtaining AI training data to develop many large language models. It analyses data scraping techniques, key stakeholders, and worldwide legal and regulatory responses. Finally, it provides preliminary considerations and potential policy approaches to help guide policymakers in navigating these issues, ensuring that AI’s innovative potential is unleashed while protecting intellectual property and other rights.