Helen Yannakoudakis

Helen Yannakoudakis is an Assistant Professor at King's College London and Affiliated Staff at the University of Cambridge. She is also a Turing Fellow and a Fellow of the Higher Education Academy.

Helen is working on machine learning for natural language processing with a focus on few-shot learning, lifelong learning, multilingual NLP, and societal and health applications.

Helen's work has been deployed under the Cambridge brand (Write & Improve), and has been published in leading venues in the field such as NeurIPS and ACL.

She has received funding awards from industry and academia, has served as a keynote speaker and a panelist, and has won international competitions such as the NeurIPS 2020 Hateful Memes Challenge.

Among others, she has been invited for spotlight interviews (e.g., DrivenData) and comments by media channels such as Reuters and TechCrunch.Recently, she was invited to stay at Windsor Castle to talk about AI in a two-day consultation on threats and opportunities.

Abstract

Large Language Models in Language Teaching and Assessment

In this talk, we focus on the potential for integrating large language models (LLMs) into AI-powered language teaching and assessment systems. We explore various research areas including content creation, automated grading, and grammatical error correction, while also addressing the risks and ethical concerns surrounding the use of generative AI in language learning technology. We highlight the need for further research to better understand the strengths and limitations of LLMs and to address foreseeable risks such as misinformation and harmful bias, and explore several directions for future work.