Generative Artificial Intelligence (AI) has made headlines across the world that proclaim an existential threat to humanity, with ChatGPT being heralded as the beginning of a new era in technology. This scholarship project explores how this disruption could potentially lead to enhanced teaching and learning, and skills development for Open University students by harnessing the benefits and identifying the risks for student of generative AI.
Concerns over the penetration of AI – and specifically ChatGPT – into Higher Education have focussed on plagiarism or biases in the algorithms producing distorted results (UNESCO, 2023). ChatGPT’s writing style and capacity to mimic particular styles or write at different levels of technical ability on command afford a more emotional or human quality to the writing compared to preceding AIs and are supposedly less likely to be detected for this reason. Yet the adoption of AIs by students to write their essays may result in poor levels of understanding by the learner or indeed the spread of false information or misinformation by AI that results in misunderstanding altogether.
This project aims to preserve the essay as a unit of assessment in FASS disciplines by exploring ways in which AI can be incorporated into the process of teaching and learning itself. As often isolated learners with little direct supervision, Open University students may be at vulnerable to using software such as ChatGPT and the risks outlined above. Rather than just warn students of the risks of AI which may have limited effect, the project explores how students may use AI, in the process of writing essay questions, with critical awareness of the risks as well as drawing the benefits of a dialectical learning partner or ‘study buddy’ in a constructivist pedagogy. The process of asking AI to write an essay will be the site of learning itself as the student is increasingly challenged to provide relevant information to the software thereby developing their own understanding.
The main part of this project will ask students working in groups to devise an answer for a short assignment question using various resources including AI. Each iteration requires further detail and specification from the user, including details and ideas held behind the Open University’s login screens that can not be accessed by ChatGPT. The process of getting to a near-suitable answer therefore requires increasing knowledge of theory and detail as well as skills in selecting appropriate key words to construct a suitable question for ChatGPT and evaluation of the generated answer. It is this process of putting together the appropriate question that the research will explore as a site of learning. The study takes a ‘connectivist’ (Stacey, 2013) approach of viewing learner knowledge as constructed in the connections being made between different resources, the synthesis and evaluation of knowledge in an age where information is readily available but requires skills in separating truth from falsehoods.