Skip to main content Skip to main navigation

Publikation

German Voter Personas Can Radicalize LLM Chatbots via the Echo Chamber Effect

Maximilian Bleick; Nils Feldhus; Aljoscha Burchardt; Sebastian Möller
In: Saad Mahamood; Nguyen Le Minh; Daphne Ippolito (Hrsg.). Proceedings of the 17th International Natural Language Generation Conference. International Natural Language Generation Conference (INLG), Tokyo, Japan, Pages 153-164, Association for Computational Linguistics, 2024.

Zusammenfassung

We investigate the impact of LLMs on political discourse with a particular focus on the influence of generated personas on model responses. We find an echo chamber effect from LLM chatbots when provided with German-language biographical information of politicians and voters in German politics, leading to sycophantic responses and the reinforcement of existing political biases. Findings reveal that personas of certain political party, such as those of the `Alternative für Deutschland' party, exert a stronger influence on LLMs, potentially amplifying extremist views. Unlike prior studies, we cannot corroborate a tendency for larger models to exert stronger sycophantic behaviour. We propose that further development should aim at reducing sycophantic behaviour in LLMs across all sizes and diversifying language capabilities in LLMs to enhance inclusivity.

Weitere Links