Harnessing the Power of GenAI in Sustainability Education - Reflective report

Here is a reflective report on my 6-week summer research period where I explored the potential implementation of GenAI tools to enhance sustainability education in higher education institutions.
Like

Share this post

Choose a social network to share with, or copy the URL to share elsewhere

This is a representation of how your post may appear on social media. The actual post will vary between social networks

Grab that notebook and pen you have lying around for the unusual occasion you have to hand write something. Now, draw a Venn diagram with one side titled GenAI and the other sustainability. Write whatever comes to your mind about the two topics and in the middle any connections. Now I’ve got you thinking, let’s begin! You’ve most likely heard of GenAI, in fact you’re probably tired of hearing about it, but you probably use it in one way or another. Sustainability, another term you’re likely tired of, but are you actually engaging with it? 

GenAI is, in broad terms, a technology that is programmed to generate new content based on learned language patterns and context. Models are typically trained on vast amounts of data, including webpages and online media. Outputs can vary from text, to images, to music, and code. For this research, of concern were models like GPT (Generative Pre-trained Transformer), which generate human-like text responses. The development of AI is both rapid and complex, offering exciting uses. For example, GenAI has sped up drug discovery by accurately predicting protein structures (Jumper, J. et al, 2016) and has carried out the task of a personal tutor for students (Luckin, R., et al, 2021). But there are challenges too: AI can reflect biases from its training data (Bender, E.M., et al, 2021) and requires a lot of energy to run, raising concerns about its environmental impact. 

While knowledge and capabilities of GenAI have been advancing rapidly, in the meantime we have reached a critical tipping point in the climate change trajectory. With a rising human population and increasing threat to natural resources due to human activity, we face a “battle for our lives” (UN Secretary-General António Guterres). Tackling this challenge requires coordinated global action across all sectors. Higher education institutions (HEI’s) are uniquely situated in this battle, with their potential to deliver sustainability education and empower students to lead the transformations needed for a sustainable future.  

Over the past 6 weeks, I have explored the perspectives of GenAI tools to realise the role this technology can play within HEI’s to increase focus on sustainability. This is a qualitative study, where I employed the method of ‘thing ethnography’, with GenAI chatbots as the ‘thing’s. In collaboration with the sustainability service at the university of Leeds, I organised a focus-group with staff members from the Leeds University Network for Sustainability in Higher Education (LUNSHE). The aim of this focus-group was to facilitate discussions to help critically analyse and validate GenAI’s suggestions. 

To kick-off this research I took a dive into the literature concerning (i) GenAI, (ii) thing ethnography, and (iii) Sustainability education. As you can probably imagine, there is no lack of literature covering these topics! With an exhaustive library of publications potentially relevant to my interests (20,000+ results from google scholar searches), I needed to narrow my view. Thus, my literature review commenced with a look into (i) GenAI in education, (ii) Education for Sustainable Development (ESD) frameworks, and (iii) thing ethnography with AI.  

To gain GenAI tools’ perspectives, I conducted a set of semi-structured interviews with ChatGPT, Copilot, Claude, and Gemini. I designed a list of prompts for the chatbots with a focus on addressing education for sustainable development (ESD) competencies outlined in two frameworks, (Advance HE, 2024, Framework for ESD) and (UNESCO, 2020, ESD: a roadmap). From the interviews, I recorded common responses to gather the perspective of the ‘thing’. I presented a summary of GenAI’s responses along with the prompts to 3 groups of educators from the university during a focus group. Each group was assigned the task of discussing responses to each of the prompts. Discussions were recorded and transcribed. I then conducted a thematic analysis to distil findings. We will finally prepare informed guidelines on where GenAI can be used to support ESD within HE, along with potential challenges. 

Interviewing is a technique I haven’t practiced since studying psychology A-Level in school. I was reminded of this experience whilst designing questions for the interviews with GenAI. The interview process also felt familiar, however, the nature of interviewing a non-human ‘thing’ felt somewhat unnatural to me. Albeit, I found the chatbots ability to answer my questions impressive. I found the flow of conversation with the chatbots was less natural and progressive than conversations with humans, although I did find that it was quicker to get to the point of the question, with fewer side tracks. I believe that with a good understanding of the topic of conversation, interviewing GenAI could be an innovative way to answer questions that require multiple inputs to be presented in a short, summative form. 

After interviewing GenAI, I assimilated an extensive list of innovative suggestions. Suggestions included, “GenAI can analyse student learning preferences to create tailored lesson plans that are sustainability-focussed”, “GenAI can benchmark curriculum proposals against leading examples and generate recommendations to enhance the integration of ESD”, “GenAI can continually update problems with real-time data”, GenAI can help students explore the consequences of their actions on the environment through immersive simulations”. This list barely scrapes the surface of how GenAI perceives it can enhance sustainability education. What this list does highlight however, is GenAI’s tendency to generate ‘flowery’ responses that sound really good, but somewhat lack practicality. These responses do not include the ‘how’, for example, how will GenAI select leading examples? This problem was one thing highlighted during the focus group that followed. 

The purpose of the focus group was to pull apart and critically analyse GenAI’s suggestions, which is certainly what I observed. I was fully engaged while listening to the in-depth discussions between the groups, highlighting points that I hadn’t previously considered. Although not as extensive as suggested by GenAI, the educators did agree on ways in which GenAI could be really helpful within their disciplines to help them to deliver sustainability education. The figure below highlights the agreed applications for GenAI that were common from each group’s discussions. 

 

Figure 1: GA, summary of recommendations for GenAI in ESD following focus-group discussions. 

As we have uncovered, GenAI certainly offers exciting possibilities for sustainability education, however this is not without its challenges. It was a hot topic amongst all groups that GenAI holds inherent bias from the data it is trained on. Currently, GenAI models are developed by private companies, notably, coming from first-world countries. Therefore, their outputs reflect biases, limiting their global applicability. Discussions stressed that global inputs are needed to truly enhance sustainability education, which requires international collaboration. Considering this challenge, participants suggested that universities could collaborate internationally to design customised GenAI models fit for academic and sustainability purpose.  

Another challenge highlighted regards the lack of access GenAI has to up-to-date data. As new research and data comes to light, GenAI may lag behind with its sources for some time, so the groups were sceptical on its ability to “update problems with real-time data.” Further, the participants argued that the user must already have a good understanding of the topic in question for responsible use of GenAI’s responses. GenAI has a tendency to output text that sounds good but is not always accurate or trueIt is a concern, therefore, that misinformation may be spread if people try to use GenAI for learning novel concepts. 

The findings of this research hold significant importance, as they highlight the role GenAI can play within HEI’s to enhance sustainability education. I view the challenges not as a limitation to GenAI’s role in sustainability education, but as crucial knowledge to inform the most comprehensive guidelines that encourage responsible and effective use of the tools. The nature of discussing the ethnographic study findings with academic experts has given me confidence in my quest to provide educators with guidelines that will support them to adopt GenAI tools in their work to enhance sustainability education, regardless of their primary discipline. There is potential for these applications to have immediate impact, which is exactly what is needed to drive us towards a sustainable future, encouraging me to apply this research and share what I have found not only within my home institution, but with institutions globally. 

On top, an exciting part of this summer for me was attending the 2024 edition of UNESCO’s digital learning week. This event provided me with an invaluable opportunity to explore the rapidly evolving intersection of AI and education and witness global efforts to reshape the future of education through technology.  

Throughout the week, the event focused on how AI is transforming education systems worldwide and how this transformation must be approached with fundamental values centred on human rights, social justice, and sustainability. I was one of over 900 participants from more than 100 countries gathering in plenary and breakout sessions, where policies, regulations, and best practices for a more human-centred, climate-conscious approach to AI in education were shared. The event gave way to the release of UNESCO’s AI competency frameworks for both students and teachers, which has informed my analysis of how GenAI can facilitate sustainability-focused learning. 

Prior to this event, I was feeling somewhat bogged down and unmotivated after weeks of working alone at my desk. However, the collective commitment of the global community throughout this event highlighted the shared vision of using AI to craft education systems that are not only technologically advanced but also sustainably focused and socially just, opening my eyes to the significance of my research and motivating me to press forwards. 

The concluding remarks by Mr. Sobhi Tawil, Director of the Future of Learning and Innovation Team at UNESCO, encapsulated both the event’s overarching message, and my motivation for involving myself in this crucial area of research: the future of education lies in leveraging AI and digital tools, but always through the lens of human rights, inclusivity, and sustainability. 

Moving forward, the key initiatives and frameworks introduced during this week have served as foundational elements in my research, helping me to better understand how GenAI can be harnessed to assist in the creation of educational practices that are grounded in sustainability values. 

What’s more, I made meaningful connections with leaders in education worldwide that I hope to carry with me throughout my progression. I was praised on my leadership skills when I arranged for a group of us to go out for dinner one evening. They commented “Your ability to arrange for a group of adults who you have just met to meet together in one place is amazing.” Despite sounding small on paper, I was proud of myself for making such an impression on these people. 

Two women standing in front of a blue backdrop 

Figure 2: Myself and supervisor, Vasiliki, at the UNESCO Digital Learning Week 2024, 02/09/2024. 

Throughout this summer, I have made significant strides in my skills development and have begun to lay the groundworks for future prospects. In collaboration with my supervisors, I am currently drafting an article based on this project, with the aim of publishing it early next year. This experience has not only deepened my research skills but also sharpened my academic writing abilities. 

I have also discussed the possibility of presenting my findings at the University of Leeds annual Sustainability Conference. The organisers have expressed strong interest in having my research featured, which would provide an excellent platform to share my work with a broader academic and professional audience. 

In addition, I have connected with scholars from other universities who are working on related topics. I hope that these collaborations will allow us to cross-reference our findings, ultimately strengthening the conclusions of our respective studies. By fostering these academic relationships, I hope to enhance the impact of our research and develop more comprehensive guidelines. 

What’s more, during my participation in UNESCO's Digital Learning Week, I spoke about my research with several attendees, who encouraged me to apply to present at next year’s edition of the event. This feedback has reinforced the ambition I have to expand the reach of my work and engage with a global community of researchers. The prospect of presenting at such a prestige, global event excites me as I endeavour to form global partnerships and see this research benefit institutions world-wide. 

Moreover, I have learnt the importance of celebrating each step of my journey, and the necessity of maintaining a healthy balance between my responsibilities and well-being. This experience has prompted me to reflect on my tendency to overwork, often neglecting my wellbeing as a result. I have realised that I can work far more effectively when I allow myself time to rest and recover, and that pausing is not a sign of weakness, but a demonstration of self-discipline and knowing one’s limits. I have always viewed my eagerness to seize every opportunity as a strength; however, I now recognise that this quality of mine needs some refining to prevent burnout.  

In summary of my first summer as a Laidlaw scholar, I feel I have gained a deeper understanding of myself and my abilities. Heading into my third year of university, I feel more confident than ever in my ability to maintain a healthy balance between work, study, and leisure as I embark on another exciting and challenging year. I believe I have developed crucial leadership skills that will help me guide the competition climbing team to success in my role as Competition Secretary for the Climbing Society. I am also eager to continue my journey as a Laidlaw Scholar, and I look forward to advocating for this incredible opportunity. 

References 

Advance HE (2024), Framework for Education for Sustainable Development, available at: https://www.advance-he.ac.uk/knowledge-hub/framework-education-sustainable-development, accessed 09/08/2024. 

Bender, E.M.; Gebru, T.; McMillan-Major, A.; Scmitchell, S. (2021). "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?" In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency. 

Jumper, J., Evans, R., Pritzel, A. et al. (2021). Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589. 

Luckin, Rose; Holmes, Wayne; Griffiths, Mark and Forcier, Laurie B. (2016). Intelligence Unleashed: An argument for AI in Education. Open Ideas; Pearson Education, London. 

UNESCO (2020), Education for sustainable development: a roadmap, available at: https://unesdoc.unesco.org/ark:/48223/pf0000374802, accessed 09/08.2024. 

 

Supervisor comments

Grace has been an excellent student to collaborate with. She has shown leadership, commitment, innovative thinking and willingness to explore an emerging area of research that may help advance teaching and learning in the HE sector. Grace’s contributions to this project over the past three months are very commendable. In addition to managing several commitments during the summer, she quickly adapted to a completely new research area and developed a range of valuable research skills. After initial guidance, Grace was able to manage her literature review and identify publications to help her develop a framework for the GenAI tools interviews, and then synthesise findings and develop the themes around which the focus group discussions would revolve. This involved learning about GenAI tools, sustainability education frameworks, qualitative methods (i.e., interviews and focus groups), ethical considerations, thematic analysis, and others. Towards the end of the project, Grace run a focus group with members from the Leeds University Network for Sustainability in Higher Education (LUNSHE), showing great collaboration and communication skills by engaging with a wide range of stakeholders. She has also demonstrated strong data analysis skills and produced a high-quality poster to communicate her findings. She has expressed interest in co-authoring an academic paper with us based on her research findings, further developing her expertise in qualitative data analysis, academic writing, and publishing. Her participation at the UNESCO digital learning week, helped her speak comfortably about her research and combine new ideas from the various presentations to understand where her research would be of value. We believe some of the findings from this project could contribute to developing guidelines for a range of stakeholders on the use of GenAI tools for sustainability education, supporting the Sustainable Curriculum principle of the University of Leeds’ Climate Plan.

Please sign in

If you are a registered user on Laidlaw Scholars Network, please sign in