That AI applications interfere with democratic processes is no longer a future topic, but has been unfolding at least since 2024, when more democratic elections than ever before took place at once, and AI made a major leap in its development and accessibility to the public.
The core panel was designed to join perspectives from both sides of the Northern Atlantic and Europe with perspectives from countries of the Global South:
- Dr. Heather Ashby, Principal consultant with Corner Alliance, working with the US Department of Homeland Security’s Science and Technology Directorate
- Dr. Hasnain Bokhari, Head of Digital Policy and Artificial Intelligence, Willy Brandt School of Public Policy
- Tabea Wilke, Trust & Safety Teaching Consortium, Stanford University, and founder of Twincler
- Moderated by Prof. Dr. Thorsten Thiel, Professor for Democracy Promotion and Digital Policies, Universität Erfurt
The expertise on the panel was enriched through a fishbowl format through which the audience could directly join the discussion.
Bringing the several very different perspectives together showed that the question of this dialog is not only one about technology, but rather also about regulatory environments, political systems and plainly infrastructure. For example, the European Union has a much more regulated environment for the use of AI, mainly guided by concerns of data privacy and consumer protection, which makes it less popular with tech entrepreneurs. One aspect regarding the US in turn is the decentralized legislation on many aspects of the use AI, which differs from state to state. For some countries in the Global South, however, the necessary computing capacity and a reliable electricity grid might already be a challenge for the development of own AI technologies.
One aspect that is often neglected in European debates is that the current focus of competition is not Transatlantic, but rather between the US and China. As AI is never just a technology, an algorithm, but always based on the data it is being trained with, the future might also show some regionalization in AI models to account for different contexts.
The debate brought up a range of examples of AI being used to manipulate electoral campaigns, from amongst others Mexico and Indonesia to the United States. And while “manipulation” was mostly meant against democratic standards, a case was brought up where a regime’s attempts to silence their opposition could be circumvented through the help of AI. There were also instances where voter information had been improved with the help of such technologies.
While these phenomena have clearly already arrived and will gain even more traction in coming years, visions of more democratic participation and consensus building through large language models, such as Google’s Habermas Machine, are still in their very early stages.
The Erfurt Transatlantic Dialog is a cooperation between the Konrad Adenauer Foundation Thuringia, Consulate General of the Unites States of America in Leipzig and the Willy Brandt School of Public Policy. We thank our partners and all participants for a vivid discussion.