Generative AI is built on large language models (LLMs) which in turn are trained on large and diverse pools of data. Researchers, corporations, and individuals rely on these LLMs to answer questions, make predictions, and solve complex problems. But that data is not governed effectively; policymakers struggle to enforce copyright and personal data protection. Moreover, content creators lack protection for their ideas and opinions. Finally, the models may be built on incomplete, inaccurate and unrepresentative data. The organizers of the conference aim to:

• Identify gaps in data governance for LLMs
• Suggest and discuss ideas to address these gaps
• Promote greater understanding of data governance as a tool for AI governance.

The conference will feature two days of panel discussions, focusing on how firms acquire data, whether firms choose to make their LLMS open, partially open, or closed to outside review, and the implications of these choices for democracy, human rights, and trust. We will explore new ideas for how to govern the data underpinning generative AI while promoting broader understanding and engagement in the governance of data. Additionally, we will examine the actions governments are taking to bridge data governance gaps. The conference will culminate with a collaborative discussion among all participants to suggest a path forward. It is free and open to all interested in discussing the data that is used to build generative AI systems.