Generative AI for critical engagement with the literature
This project proposes to explore the potential of generative AI tools for student engagement with sources ranging from academic literature to industry reports and media stories. Across different Departments and Faculties at King’s College London, critical engagement with sources and literature is one of the first steps when it comes to learning. And yet, this presents a problem for many students due to a range of problems such as language issues, limited or inadequate previous knowledge, the overwhelming volume of material, and complex terminology. Generative AI tools, if used appropriately, give students an opportunity to overcome some of these barriers to learning. A few recent publications* have outlined some of the ways students have been using generative AI tools: summarising texts, concept checking/explanation, text translation, quiz creation or question bank generation, and creating jargon lists or glossaries.
However, the use of generative AI tools comes with challenges. It has been recognised that they have a tendency to produce plausible but false information, so-called hallucinations. AI in general can produce output that is culturally, racially and politically biased, and there are security and privacy concerns. Moreover, there are dangers for generative AI tools to increase digital divide among students and this project aims at conducting primary research with the aim of generating knowledge that could help provide guidance for inclusive and equitable use of these tools. Our student cohorts are diverse in terms of gender, neurodiversity, nationality, race, ability, and previous educational background.
Therefore, this project proposes to conduct primary research with our students. First, the aim is to conduct a survey on current use among our undergraduate and postgraduate students. The second step would be to conduct focus groups with diverse groups of students where they would be encouraged to experiment with the tools with appropriate guidance in terms of prompting strategies and tips for scrutinising AI responses for example. Following the analysis of the data, a guidance and toolkit would be created for students and teachers. In addition, the output would be an academic article.
In terms of evaluation, the aim is to evaluate the conduct of the project as well as the final result. Therefore, the first two focus groups would be surveyed with the aim of improving research. Finally, there would be an additional focus group with the students not previously involved in order to test the final toolkit produced.
Generative AI tools have attracted plenty of attention and there are various assumptions floating around popular and academic conversations. Conducting primary research with students is key for us to be able to help our students develop AI literacy and use these tools as complement to their learning.