In my presentation I will first outline some hopes and concerns from the use of LLMs for democracy. I will then explain what role the emerging digital framework from Brussels can have in realising the potential, and minimalizing the risks from LLMs for democracy, beginning with the making of the AI Act itself. While the AI Act is currently being celebrated as the first comprehensive legislator framework for AI, a less convenient truth is that it were the rules about LLMS and the political powerplay around those rules that almost caused this democratic process to collapse.
Speaker: Natali Helberger (AI, media and democracy lab, UvA)
