Human-Centric AI & Security Seminar

Bias in LLMs: Measurement, Control, and Risks

Speaker: Gianluca Demartini

Pricing

• Free

Dates and Times

Contact

Tina Wu

More information

Abstract:

AI can now replace humans in decision making. What are the risks of delegating decisions to AI? In this talk we will start by discussing how we can measure and control bias in Large Language Models. We will see different examples of bias in AI and in data, and how this may be used to steer political election results. Then, we will look at how we can make people aware of existing bias and at the risks of involving LLMs in the loop when humans are in charge of making decisions. Finally, we will look at results on how people trust AI and we will discuss about how we can build responsible human-AI collaborative system that work well.

Bio:

Gianluca Demartini is a Professor in Data Science and an ARC Future Fellow at the School of Electrical Engineering and Computer Science at the University of Queensland, Australia. He is also a Dieter Schwarz Fellow at the Technical University of Munich, Germany. His main research interests in Data Science include Information Retrieval and Responsible Artificial Intelligence.