Our CLT EduTalk in March 2023 focused on the current ‘hot topic’ in education – ChatGPT and the impact of generative artificial intelligence (AI). The CLT was pleased to welcome over 85 staff and students from across campus to the online event on 7 March, where colleagues from the Department for Health, the Department of Computer Science and the School of Management gave some excellent insights into the ways in which they have been working with generative AI in their teaching and scholarship. The talks below refer to ChatGPT-3.5.
Case Study 1:
Evaluation of ChatGPT Capabilities
Joshua Evans, Department of Computer Science
Joshua gave an brief introduction to ChatGPT and outlined its’ capabilities. He has been looking at generative AI tools and considering them from the angle of academic integrity. Most recently, his focus has been to specifically evaluate ChatGPT’s performance at generating assessment responses.
Introduction to ChatGPT and its Capabilities
Case Study 2:
Assignment audit using ChatGPT
James Fern, Department for Health
James has been carrying out an initial audit of a selection of current UG assignment questions/briefs. In his audit, he has been looking across the course years to identify how well ChatGPT answers the assessment questions posed at different levels.
Case Study 3:
Harnessing ChatGPT as a tool to aid critical thinking
Kim Watts & Dr Carl-Philip Ahlbom, School of Management
Carl-Philip and Kim have been piloting use of ChatGPT as a tool to support the development of students’ higher order thinking skills (their ability to analyse, evaluate and synthesise) as part of their assessment. They have been engaging students in considering the limitations of the tool and the importance of promoting and harnessing human critical thinking.
Case Study 4:
ChatGPT – Reliance on Linguistic Form
Dr Harish Madabushi, Department of Computer Science
An extract from the work being carried out by Harish Madabushi in Computer Science, looking more closely at how ChatGPT functions.
ChatGPTs Reliance on Linguistic Form in Inferential Reasoning