Julian Gierenz

Hi! I’m a PhD Candidate and Research Associate based at the Chair for Political Science, especially Digital Transformation at the University of Bamberg. At the heart of my work lies the critical question of how artificial intelligence, particularly Large Language Models (LLMs), are reshaping political communication and governance. I investigate this technology’s dual-edged nature: exploring its potential to strengthen democratic processes like deliberation on one hand, while critically analyzing its deployment for control and manipulation within authoritarian regimes on the other. This dual focus stems from a commitment to understanding how technology can be steered towards supporting, rather than undermining, democratic values.

My Doctoral Research: Leveraging AI for Scalable, Quality Deliberation

We see technology increasingly exploited in ways that can undermine democratic values. My motivation is to actively counterbalance this trend. I engage deeply with technology, asking how we can strategically leverage it to strengthen democratic institutions and norms, aiming for my research to contribute towards a more democratic, just, and fair world through the thoughtful application of technology.

My PhD focuses on small-group deliberation, a cornerstone of democratic systems. These interactions are fundamentally language-driven, guided either by the group itself or specialized moderators, always striving for discussion grounded in deliberative norms. However, a critical limitation persists: the challenge of scaling these valuable processes without sacrificing quality as participant numbers grow.

Until recently, complex language-based interaction was considered an exclusively human domain. However, the emergent capabilities of Large Language Models (LLMs) reveal their surprising ability to grasp fine-grained, context-dependent nuances – a fundamentally relevant characteristic for guiding deliberation. This leads to the central questions of my dissertation: Can LLMs be utilized to help guide or manage these language-driven processes, at least in part? Could LLMs, tailored for this niche, be the missing piece in efforts to automate aspects of group deliberation, thereby enabling it to scale and potentially resolve the classic dilemma between broad participation and high deliberative quality? Exploring these questions, bridging democratic theory with computational techniques like Natural Language Processing (NLP), is the core of my doctoral work.

Research Associate: The Flip Side – AI Control in Authoritarian Contexts

Complementing my doctoral research on the potential of LLMs to enhance democratic processes, my work as a Research Associate examines the crucial flip side: how these same powerful technologies are actively shaped, controlled, and potentially instrumentalized within authoritarian contexts. Contributing to the BIDT-funded project ‘Authoritarian AI’, I analyze the mechanisms of strategic control, such as state-sponsored development and censorship, that regimes employ over LLMs. We investigate how state-aligned or biased models become embedded within societies like Russia and assess the spillover risks of authoritarian data influencing broader AI ecosystems. Understanding these dynamics of control and manipulation provides a critical counterpoint to the exploration of AI’s pro-democratic applications.

Bridging Disciplines with Computational Methods

My commitment to using Computational Social Science, particularly Natural Language Processing (NLP), stems from the conviction that effectively analyzing AI’s role in contemporary politics requires engaging directly with the technology itself. These methods provide the essential toolkit for systematically examining complex communication patterns and model behaviors at scale. While the field is rapidly evolving and demands continuous learning, developing proficiency with these tools is crucial for dissecting the subtle dynamics of productive deliberation and for identifying patterns of bias or control within AI deployed across different political regimes. This ongoing engagement, driven by a deep desire to understand the inner workings of these influential technologies, enables me to pursue deeper, evidence-based insights into AI’s multifaceted political impact.

Research Interests Include:

• Computational Social Science & Natural Language Processing: Applying and adapting methods to analyze political communication and AI behavior.
• Artificial Intelligence & Politics:
  • AI for Democratic Innovation: Exploring uses in deliberation scaling, facilitation, and enhancing participation.
  • AI Governance, Ethics & Bias: Examining challenges across both democratic and authoritarian contexts.
  • Authoritarian AI: Investigating state control, censorship, and information manipulation via AI/LLMs.
• Democratic Theory & Practice:
  • Deliberative Democracy: Focusing on theory, epistemic quality, and scaling challenges.
  • Social Epistemology: Understanding group justification and collective reasoning processes.
• Digital Politics & Information Environments: Analyzing the broader impact of technology on political life.

Let’s Connect

I’m always eager to connect with fellow researchers, practitioners, and anyone exploring the complex intersection of artificial intelligence, political science, and communication. Please feel free to reach out if you share these interests or want to discuss these rapidly evolving topics!