Responsible AI Research Lab at The University of Queensland, Australia
Building AI for public good.
DLab is an interdisciplinary research group exploring how data, people, and AI affect each other.
Focus
Responsible AI and sociotechnical design
Approach
Interdisciplinary collaboration with real-world partners
Community
Researchers, students, and institutions shaping the future of AI
Perspective-Aware LLM Evaluation
We analyze how large language models encode political and demographic perspectives, with a focus on bias, relevance judgments, and summarization effects.
Human-Centered Data Quality
Our research combines metadata, crowdsourcing, and information retrieval methods to improve quality assessment and trust in unstructured data workflows.
Social Media Integrity and Safety
We study misinformation, harmful content, and online persuasion, developing methods and tools for fairer and more transparent sociotechnical systems.
Latest Publications
The 5 most recent entries from our publications page, highlighting current work across LLMs, fairness, relevance, and social media analysis.
Pietro Bernardelle, Samaneh Mohtadi, Stefano Civelli, Joel Mackenzie, and Gianluca Demartini. LLMs Encode Relevance as a Layer-Wise Cross-Lingual Signal. In: The 35th International ACM Conference on Knowledge and Information Management (CIKM 2026). Rome, Italy, November 2026.
Samaneh Mohtadi, Pietro Bernardelle, Joel Mackenzie, and Gianluca Demartini. Persona Conditioning as an Assessor-Sensitivity Probe for LLM-Based IR Evaluation. In: The 35th International ACM Conference on Knowledge and Information Management (CIKM 2026). Rome, Italy, November 2026.
Martin Paul Wessel, Timo Spinde, Jürgen Pfeffer, and Gianluca Demartini. Definitional Sensitivity in Media Bias Detection: A Multi-Definition Dataset and Benchmark. In: The 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP 2026) - Findings paper. Budapest, Hungary, October 2026.
Bing Tuo, Tim Miller, and Gianluca Demartini. Easy to Read, Easy to Trust: How Processing Fluency in LLM Explanations Drives Over-Reliance in Hate Speech Moderation. In: The 2026 ACM Conference on Human-AI Complementarity and Alignment (HCOMP 2026). Alexandria, VA, USA, September 2026.
Pietro Bernardelle, Stefano Civelli, Leon Fröhling, Riccardo Lunardi, Kevin Roitero, Gianluca Demartini. Political Ideology Shifts in Large Language Models. In: Information Processing & Management (IPM), Elsevier. August 2026.