Trolls versus democracy? How Artificial Intelligence impacts political campaigns

In this “super election year”, where more than 50 per cent of the world’s population will be going to the polls, Artificial Intelligence (AI) is poised to alter the relationship between parties and voters. In their new research chapter “KI und datengesteuerte Kampagnen: Eine Diskussion der Rolle generativer KI im politischen Wahlkampf” (AI and data-driven campaigns: A discussion of the role of generative AI in election campaigns), Dean of Executive Education Andrea Römmele and Centre for Digital Governance postdoctoral researcher Dr Philipp Darius provide a discussion on the potential dangers, risks and benefits of generative AI. 

AI can be both constructive and destructive to political campaigns

Generative AI is the use of AI to create images, texts, songs and other content which will be disseminated via social media channels or websites. “Generative AI can be a powerful tool for political campaigns […] by generating text-based and visual content and providing support for data analyses,” the Professor of Communication in Politics and Civil Society and the postdoctoral researcher write. “However, it also increases the existing risks of digital media as political communication environments if political actors do not adhere to democratic norms and values.” 

According to the scholars, no central algorithm or AI model is able to manage a campaign’s budget, messages and resources yet. Moreover, compared to the traditional tools of campaigning such as surveys or focus groups, AI performs similar or worse, they point out. However, “new technologies such as generative AI can accelerate the division of political campaigns into scientific and subversive campaigns”, the authors write. This means that the communication voters experience could be constructive or destructive to a campaign’s message. Examples of the latter include 1) deep fakes: AI-generated images that appear real, but do not show scenarios that actually happened; 2) hashjacking: the reframing of a campaign’s hashtags; and 3) astroturfing: the organised creation of fake social media accounts to influence a given campaign.

State and civil society supervision help ensure strong campaigns and strong democracy

How can we respond to these dangers and leverage AI’s potentials? Darius and Römmele stress that 

• online campaigns need to be adequately monitored; 
• the press plays a vital role in evaluating and fact-checking information; and 
• citizens need to receive digital education to navigate this new landscape.

With the emergence of the EU’s Digital Services Act, the authors also point to the role of the Digital Services Coordinator (DSC) – national governmental bodies overseeing digital platforms and services.

The DSC, in Germany the Bundesnetzagentur, needs to build up knowledge, critically examine campaigning actors, and delete inauthentic or harmful content. Nongovernmental bodies can contribute to this encompassing mission with their analysis and criticism, too, they write. Thus, Darius and Römmele recommend building national task forces in which civil society, state actors and platforms work together to monitor election discourses. These groups of actors need to commonly identify inauthentic behaviour, false information and disinformation campaigns by external actors to work towards safe and fair digital environments for democratic elections.

The chapter appeared in the Festschrift Informationsflüsse, Wahlen und Demokratie (Information flows, elections and democracy), edited by Thorsten Faas, Sascha Huber, Mona Krewel and Sigrid Roßteutscher. The Festschrift was published in honour of Rüdiger Schmitt-Beck, a leading German political communication scholar.

Academic reference
Darius, Philipp, and Andrea Römmele. “KI und datengesteuerte Kampagnen: Eine Diskussion der Rolle generativer KI im politischen Wahlkampf,” 199–212. Nomos Verlagsgesellschaft mbH & Co. KG, 2023. doi.org/10.5771/9783748915553-199.