Since its release at the end of 2022, we have seen widespread uptake of, and engagement with, ChatGPT (the free version) and GPT-4 (included in the updated and subscription version) in many sectors. It was clear right from the start that this powerful chatbot will have far-reaching implications for education in general and for higher education in particular. Some commentators from within the university sector went so far as to say that the new generation of “Generative AI” (of which ChatGPT forms one exponent) will necessarily lead to a fundamental rethinking of the nature of learning and the way we engage with knowledge, not only in learning programmes, but also in professions and in work sectors.
Both students and lecturers are currently experimenting with the various Generative AI chatbots (and other applications) that are becoming available. Many opinion pieces and articles on specific aspects of matters regarding teaching, learning and assessment are appearing in popular and professional journals.
Some universities quickly developed institutional positions on ChatGPT or internal guidelines for academics and students. Some developed internal workshops or seminars, often including an outside public. We can expect to see more activity in this space in the near future.
A recent whitepaper by a group of academics from a number of German universities provides an insightful overview of the nature of the technology in ChatGPT and related applications.1 This whitepaper was developed under the leadership of the University of Hohenheim.
The whitepaper provides an overview of where generative AI fits on the scene of digital developments, explaining the underlying information technologies and models (Artificial intelligence – machine learning – generative AI – large language models) to which ChatGPT presents a user interface as a conversational agent. It explores perspectives that can inform “the more extensive organizational sensemaking processes on embracing and enclosing large language models or related tools in higher education” (from the Executive Summary). Helpfully, it provides concrete examples of the potential of generative AI for higher education.
The paper then provides guidance for students and lecturers to make effective and appropriate use of ChatGPT and other similar tools.
- Recommendations for students include respect for good scientific practice in the use of sources; adherence to institutional regulations on assessment; reflection on how they will use ChatGPT to further their learning goals; guidelines on prompting as part of a search strategy to obtain valuable research results (which is a new skill to be developed by users of such tools); and using ChatGPT as a writing and learning partner.
- Recommendations for staff include reflection on the learning objectives that underpin their teaching; ways to boost learning through the use of ChatGPT; and possibilities for the development and personalisation of learning materials.
- Importantly, the whitepaper provides a detailed set of recommendations for lecturers on the design of assessment tasks.
- In the cases of students and staff, the whitepaper draws attention to risks associated with the use of ChatGPT, such as the reliability of sources, and the infringement of copyright, which entails that the information generated by ChatGPT needs to be verified.
In all of these categories, the guidance is structured in the form of a number of clearly formulated recommendations. Some key perspectives from the whitepaper are that ChatGPT emphasises the need for students to develop critical and structured thinking skills and exercise creative judgement in the use of AI; that higher education institutions should develop policies for the use of AI in general, and not specifically for ChatGPT; and the salient caution on the limitations of ChatGPT, including the fact sometimes it may produce unreliable results (“hallucinations”), amongst others, based on fake academic papers, although the whitepaper states that GPT-4 has reduced this risk significantly. There is also the potential for ChatGPT to reinforce existing societal biases. Finally, the underlying models of ChatGPT were primarily trained on data up to September 2021, which means that it cannot take more recent information into account.
The whitepaper concludes with some guidelines that institutions may find useful in developing appropriate policies for the use of AI, and with an outlook which argues that “the debate and the innovation should focus on the potential benefits of generative AI, such as improved learning, teaching, and the creation of equal opportunities for different groups of students” (p. 39). The whitepaper argues that AI will transform the student lifecycle, from admission to enrolment, careers services, and other aspects of higher education management. It recommends that higher education institutions should encourage broad, multi-perspective dialogue among many stakeholders in higher education on the impact of generative AI, as part of a systematic process of the digital transformation of higher education.
For the SADC region, an important takeaway from the whitepaper is how to ensure that all students have equal access to the benefits provided by this technological breakthrough.
The comprehensive view taken in the whitepaper makes it useful for a wide audience of users in higher education: lecturers, students and higher education managers alike.
H. Gimpel et al., Unlocking the Power of Generative AI Models and Systems such as GPT-4 and ChatGPT for Higher Education. A Guide for Students and Lecturers. University of Hohenheim, March 20, 2023.