Bio
Elfia Bezou-Vrakatseli is a PhD candidate in the UK Research and Innovation Centre for Doctoral Training in Safe and Trusted Artificial Intelligence at King’s College London, where she is also the representative of the Equality, Diversity & Inclusion committee. Her expertise is in argumentation and her research focuses on using argumentation tools to enhance the communication between humans and between humans and AI systems. On top of her research, Elfia is a Graduate Teaching Assistant at King's and she is also undertaking the Associateship of King’s College London programme. Prior to these, Elfia obtained an MSc cum laude in Artificial Intelligence at Utrecht University and an Integrated Master in Applied Mathematical and Physical Sciences at the National Technical University of Athens.
Research interests
Elfia's research focuses on the exploration of natural language texts, specifically of ethical debates, via tools of argumentation, to the end of supporting dialogue and debate between humans and between humans and AI systems, with respect to transparent and rational reasoning. One of the main challenges in developing safe and trustworthy AI systems, particularly with the recent advancements in large language models, is ensuring that these systems' values align with those of humans—a concept known as 'value alignment'. Elfia's research is dedicated to enable human-AI dialogue, purposed to support joint reasoning (i.e., ‘inquiry’), so as to ensure that AI decision-making is consistently aligned with human preferences, promoting a harmonious integration of AI in society.