UKZN’s School of Life Sciences (SLS) hosted a dialogue on the practical applications of Artificial Intelligence (AI) and ChatGPT in teaching and learning, and research in Science, Technology, Engineering, and Mathematics (STEM).
Dean and Head of the SLS Professor Ademola Olaniran described AI as computers and machines’ ability to perform human tasks, enabling machines to learn from experience and creating opportunities for progress in critical issues in health, education, and the environment.
ChatGPT, the AI chatbot developed by OpenAI and released in November 2022 was trained on massive text data to deliver human-like responses to prompts, engage in written conversation, and even generate original creative writing pieces and draft research abstracts, causing concern in the Higher Education sector. It made headlines for being able to pass prestigious graduate-level university examinations.
‘[According to CNN Business], this will change the way we use the Internet and the way we create. It adds to the ongoing challenging questions we have around how AI tools can upend professions and shift our relationship with technology, and enable students to cheat,’ said Olaniran.
The discussions were targeted at initiating a regional community of practice on the use of AI tools in teaching and learning, and research spaces.
Academic Leader (AL) of Teaching and Learning (T&L) in the SLS Professor Andre Vosloo chaired the panel comprising Olaniran, Professors Hafizah Cheniah and Gueguim Kana of the SLS and the University of Johannesburg’s Vice Dean of Teaching and Learning Professor Grace Leung.
‘We need everyone’s input and creativity to get around this disruption and harness the potential of these new technologies in our diverse fields,’ said Vosloo.
Cheniah, whose research focuses on antimicrobial resistance and drug discovery using microbes, is interested in using active and blended learning strategies to engage students and improve skills development. She emphasised that ChatGPT is a narrow AI that focuses on narrowly defined tasks and should not be anthropomorphised or feared.
Cheniah said it is important to formulate concrete recommendations not only for detection but also to improve students’ learning processes and teachers’ delivery.
She highlighted the limitations of the tool but said it could bring about a shift in skills-based and problem-solving educational approaches. She delved into critical issues associated with AI technology from ethical concerns to the entrenchment of the digital divide.
‘ChatGPT is this worldwide experiment that OpenAI is carrying out, using us to further train its machine tools,’ said Cheniah.
‘AI is not a replacement for human teachers; it’s a tool to augment and support our humanised, transformative teaching,’ she said.
She suggested focusing on knowledge application and student engagement with ChatGPT outputs that encourage critical thinking, creativity, and problem-solving.
Leung is a computer scientist who teaches software engineering ethics and computer forensics. She addressed the topic of ChatGPT being a co-pilot in the classroom, noting that there are several applications of this technology that could be taken advantage of. She described large language models and addressed issues including copyright, racism, inaccuracies, old knowledge, and the re-training of the model. She demonstrated ChatGPT by asking and answering multiple-choice questions, creating lesson plans, and writing academic essays, while noting its limitations.
‘If used correctly, ChatGPT can be a good tool to start someone on their lesson plan and reinforce their learning,’ said Leung.
Leung also highlighted other AI tools such as Microsoft Bing for up-to-date information.
Kana, a bioprocess technology expert with a keen interest in the application of AI in process modelling and optimisation, began exploring the use of AI a decade ago. He spoke about the future of postgraduate research and harnessing the power of ChatGPT, detailing the practical use of appropriate prompts for the entire scientific research process.
‘The capabilities of this technology are tremendous; it will outline the most important element of your research and provide information to include in your proposal: the background, the research gap, the research objectives, the methodology to be used, the expected outcome, and can even advise on sources of funding,’ said Kana.
‘ChatGPT could streamline the effort in developing research work and give enough time to our students to acquire skills that will increase their employability,’ he said.
‘There is a need to re-think not only how we assess students, but what we teach and how we teach it,’ said Kana.
In his role as AL of T&L, Vosloo has guided the SLS through a tumultuous period featuring remote online learning, AI, and chatbots through active engagement with technology.
‘The ship has sailed on whether we should embed AI in Higher Education systems generally; we cannot responsibly send students out into the world who cannot drive this technology,’ he said.
‘What we’ve discussed has primed us for the power and potential that AI and ChatGPT have, but what is also critically important are the limitations; this is not the magic bullet or fix-all; it’s an enormously powerful resource with major limitations to be aware of,’ he added.
Words: Christine Cuénod