What impact do social bots that use Large Language Models (LLMs) have on the quality of political discourse? The project investigates communicative AI in the social domain of political discourse using discourse monitoring and discourse intervention and thus with a largely experimental approach. The case studies are debates in German on the topic of climate change on X, Mastodon and Bluesky.
By combining discourse theory with the latest innovations in the field of LLM, the aim is both to observe and intervene in public political discourse. To this end, a group of public speakers on the topic of climate change will be included, who give their consent for bots to participate in the debates they initiate.
This makes it possible to examine the effectiveness of social bots in political discourse by comparing discourse patterns with and without bot intervention and analyzing them through accompanying user surveys. The project is led by Prof. Dr. Cornelius Puschmann and Dr. Gregor Wiedemann.
About the Overall Project
The project is part of the overall project “Communicative AI (ComAI) – The Automation of Social Communication“. The research group of the German Research Foundation (DFG) is investigating the consequences, risks, but also the potentials of the profound change in the media environment brought about by communicative artificial intelligence (AI). In addition to the HBI (Prof. Dr. Wiebke Loosen) and the ZeMKI (Prof. Dr. Andreas Hepp), which coordinate the research group, the University of Vienna and the University of Graz are also involved.
A total of nine research projects and one coordination project are investigating how social communication changes when communicative AI becomes part of it. The project brings together leading researchers from the fields of communication and media studies, human-computer interaction, sociology of knowledge, governance research and media law. They are united by the goal of systematically analysing the transformation of social communication under the influence of artificial intelligence by studying the consequences of its use in different social areas and the social discourse about it.
The research will focus on social pioneers, the development of interfaces, the legal and corporate handling of communicative AI, its role in journalism, in public (online) discourse, in personal everyday life through technological companions, in the health sector, and in learning and teaching.
A shared research environment for the participating institutions will be set up for the research group in order to increase the visibility of the research group’s findings for decision-makers in various areas of society across all locations. This will be accompanied by the central task of identifying possible future scenarios for the dissemination and impact assessment of automation processes at different levels of social communication.
Subprojects at the HBI
In addition to the project “Political Discourse: Communicative AI and Deliberative Quality”, two other projects are being carried out at the HBI:
- The Juridification of Communicative Artificial Intelligence: The project, led by Prof. Dr. Wolfgang Schulz, investigates the juridification of communicative AI. The focus is on the legal framework for communicative bots (in particular ChatGPT) and social bots (in particular X and Facebook) – on the one hand from the perspective of communication law, and on the other hand from the perspective of emerging AI regulation.
- Journalism: Automation of News and Journalistic Autonomy: The project, led by Prof. Dr. Wiebke Loosen, investigates communicative AI in journalism by analyzing the challenges for journalistic autonomy at the interactional, organizational and societal levels.
Further information
Further information on the project can be found on the project website https://comai.space/.
