News from Imperial: Insights from the Imperial College London Festival of Learning and Teaching

By Dr Lesa Kearney, Senior Clinical Teaching Fellow, Collaborative Partnerships Office, Imperial College London

Imperial College's Festival of Learning and Teaching is an annual event that brings together staff from all faculties to explore the latest advancements in higher education. Held from 15 to 17 May, the festival kicked off with a keynote speech from Professor Oguz Acar, Professor of Marketing and Innovation at King’s College on discovery-based learning (DBL) and highlighted the transformative potential of large language models (LLMs) in reshaping education.

As a learning approach that promotes problem construction and problem solving but without detailed learning outcomes, DBL might not feel like the most efficient way to acquire knowledge.  However, Prof Acar pointed out that skills in creativity and analytical thinking are the most sought-after skills by employers in all fields. As educators, we were encouraged to see our role as providing frameworks that give structure and rigour in this discovery process; and that we encourage students to learn through a series of "smart failures" on well-defined problems and hypotheses. This approach, Prof Acar asserts, fosters resilience, adaptability and problem-solving capabilities that are invaluable in an AI-driven future.

Discussion points amongst faculty included potential “atrophy of the student voice” if they come to rely on LLMs to generate written narratives, and of course the overall risk to academic, particularly in assessment.  These concerns were countered with recommendations that assessment should focus on higher-order skills such as verification and critical review. The Department of Computing gave an example of a new assessment that asks students to select and engage with an LLM to produce a business plan.  Part of the marking criteria includes demonstrating exactly how the LLMs are used to achieve the best possible plans, and the mitigation strategies for LLM shortfalls.  Another interesting opportunity proposed to all faculties was the design of formative assessments that incorporate LLM-generated feedback to be given directly to students.  Research indicates that learners often respond better to feedback in this type of format as it feels more transactional and less provocative, and therefore more likely to aid improvement.

A key takeaway from the festival was the need for us all to engage with AI, and for most of us that means how we interact with it, in what has become known as ‘prompt engineering’. As one speaker pointed out, “It's not necessarily AI we are in competition with, it is the humans who know how to use AI better than we do”.

The latest Guidance for Imperial Students on the use of ChatGPT emphasises that submitting work and assessments created by someone or something else, as if it was your own, is plagiarism and is a form of cheating and this includes AI-generated content.