Information can be misleading, incomplete, or completely fabricated. Being exposed to such a content influences our way of thinking and our decision-making processes thus creating risks to our safety as individuals and to the democratic process as a society.

In our research, we develop computational methods based on Artificial Intelligence (AI) to not just detect, but rather to comprehensively manage online misinformation in the broader sense of the term (i.e., identify, measure, index, surface, and adjust for it) by also involving humans.

Human-in-the-loop AI can support supervised machine learning models to be more transparent, explainable, and fair as humans can be systematically deployed to manage bias and edge cases where mistakes made by AI could otherwise have critical consequences. Our research outcomes include new datasets for misinformation detection and human-in-the-loop AI models that detect and explain why a certain statement is false. The benefits of this work are safer online information environments with less misleading content.

An additional outcome from this line of research has been the UQ Election Ad Data Dashboard which was tracking social media political ads during the Australian federal election campaign in 2022. We used AI methods to categorise and analyse ad text and images to understand their sentiment, topic, electorate targeting, and other dimensions.

Credit on this project also goes to the following research assistants:

Jaiden Harding (Dashboard) 
Ryan Harvey (Data Analysis) 
Jay Huynh (Data Infrastructure) 
Josh Marsh (Data Infrastructure and Analysis) 
Robert Nguyen (Data Analysis) 
Rudra Sawant (Data Analysis) 
Ankit Sharma (Dashboard) 

Selected articles and news relating to this project:

Project members

Associate Professor Gianluca Demartini

Associate Professor
School of Electrical Engineering and Computer Science

Dr Lei Han

Postdoctoral Research Fellow, School of Information Technology and Electrical Engineering