This article is a repost from City University of London.

In December, City’s Generative Artificial Intelligence (AI) Task and Finish Group and City Students’ Union (SU) held the first inter-society debate to explore what place AI has in education. The debate was hosted by the City Debating Society and included representatives from four student societies: UX.City, IEEE City Robotics Society, Computer Science Society and Data Science and Artificial Intelligence Society. Also joining the debate was AI itself as D.BATE V0.02 with one of the students presenting the AI responses.

The debate was opened by Christabel Carter, SU Advice Manager, who raised awareness of the academic misconduct support and guidance available from the SU. Richard Knott, Academic Learning Support Co-ordinator from Learning Enhancement and Development (LEaD), then highlighted the range of support available to students from the Academic Skills team.

Our compere for the evening was the President of the Debating Society who posed a series of questions and after each round of questions, the audience voted on which society had provided their favourite or most convincing responses. This blog post summarises the key themes from the debate.

What place does AI have in education?

All societies agreed that AI has a place in education and that it wasn’t something we could avoid; however, we need to be aware of the ethical issues and how to use it responsibly. They cited benefits for both staff and students, but it was noted that AI is not a substitute for human intelligence and should be used to assist humans – augmentation, rather than automation.

“If humans are augmented with AI and the opportunities that it would open up could be very vast, endless and it could benefit many people.” UX.City Society

Examples of where AI could benefit education included:

• Supporting staff with lesson building and enabling the production of more personalised learning/tailored content and providing a more engaging experience for students.
• Acting as a study buddy by responding to students’ questions and helping to explain content.
• Supporting neurodiverse students, in terms of the writing process or helping with managing information.
• Automation of administrative tasks, e.g. assisting with grading to make the process more consistent and reduce workload on educators.
• Rethinking how we assess knowledge and ability to use that knowledge, rather than assessing the outputs themselves, where marking might be biased because some students write better than others.

?“I did all of the work myself, but then I give it to AI to fix the English part for me.” Computer Science Society

About 75% of the students in the room had used AI for both personal and professional purposes, e.g. idea generation, breaking down difficult learning materials, checking computer code, helping with writing (especially emails and for non-native speakers of English), image generation and acting as a companion. It was suggested that students shouldn’t be ashamed of using it because it’s an openly available resource.

I’ve treated as AI as a friend and talk to it as a friend, and how it can help me as a friend.” Data Science and Artificial Intelligence Society

Will AI chat models and future AI make the current education system redundant?

“In my opinion, we need to develop testing methods that are not related to an individual’s ability to present their intelligence and knowledge on a specific subject in one specific way.” Computer Science Society

Most of the societies felt that the current educational system won’t be made redundant by AI, but it does need to change, especially assessment. It was suggested that the roles of educators and institutions may need to adapt in response to a changing landscape with the suggestion from the D.BATE V0.02 (AI participant) of a symbiotic relationship between AI and educators in the form of AI driven tools for teaching and learning; however, the societies noted that the human aspect is irreplaceable. We need to remember that education isn’t just about learning information; it’s about fostering critical thinking, creativity and interpersonal skills.

“Educators bring empathy, mentorship, and the ability to inspire a crucial dimension that AI lacks.” D.BATE V0.02 (AI Participant)

Whilst some students talked about the role of AI as a companion, it was felt that key activities such as mentoring and social connections can’t be done by AI. The whole university experience is not something that can be delivered by solely by a computer, so we need to make sure we don’t put all our trust in AI.

But in the future, we might find the current education system is made redundant as AI might have built the capacity to replace certain educational tasks and what we need to learn will change as jobs get replaced by AI. For example, will Computer Science courses still teach students how to code?

Are current chat models overrated?

“It’s like having a clever friend who occasionally misses the point. Let’s appreciate the strength while keeping in mind it’s not immune to the occasional brain freeze.” D.BATE V0.02 (AI Participant)

There were mixed thoughts on this question. It was suggested that current chat models are overrated due to limitations in the size/scope of the data that AI tools have been trained on, and that they magnify biases. Whilst AI tools can excel in some areas, we have to manage the expectation that we think it can do more than it can. For example, ChatGPT can’t play chess, neither can it write your assignments well as whilst it is good at generating text, it may have limited domain specific understanding or lack ethical considerations. There was a suggestion that we need to be able to identify the correct tool for the job, for example, the current chat models are good at communicating a message and answering questions. However, the paywall aspects of current models means they have limited benefit for many sectors and parts of society. It was suggested that despite its limitations and inconsistencies which impact current use, from a technological evolution perspective, this is one of the most impressive discoveries we have seen.

“As a layperson, I see the issues. I see why the understanding might be that it’s overrated, but from a tech perspective, the developments that have gone into ChatGPT from every discipline, from language processing, from just the broad knowledge that it has, it is potentially one of the most impressive discoveries I have ever seen.” Computer Science Society

Should there be an educational module in universities or college or high school on AI and how would it be delivered?

“There’s no running away and we cannot stop students from using AI, so might as well just teach them how to properly use it. “ IEEE Robotics Society

There was a split vote amongst the audience as to whether AI should be taught in schools. The societies taking part in the debate felt that there should be an educational module teaching students about AI in the same way we teach students about research methods. It was suggested that AI doesn’t speak English in the way we do and that we should learn to speak the language of AI. The module should include prompting (with example prompts), how AI works, limitations, ethics and biases in order to facilitate effective and responsible use. One society suggested that it should be a supplementary, optional module for students. It was noted that as AI progresses, some people will be left behind, so we need to ensure that everyone has access to learn about it and how to use it.

Wider perspectives on AI

The debate then moved away from the focus on education to consider AI in society more generally. Key themes included:

• We cannot stop innovation in AI, but governments do need to consider how we monitor and control this to avoid individuals or commercial organisations directing the direction and development. There were also concerns about AI becoming self-aware and suggestions to put in place appropriate safeguards, as well as concerns about over reliance on AI with parallels drawn to social media addiction.
• Strengths of AI, beyond use in education, were noted and it was felt that AI could help with many aspects of daily life. One example cited was a student who had used it to research good places to live in London, and subsequently lived in the nicest flat others had seen. It was also suggested that AI helps to liberate information and provide greater access to information through advances in digital translation and enabling information to be communicated in different ways.
• Limitations of AI included not being able to think for itself, so it’s basically regurgitating thoughts from humans. In addition, there was strong agreement that AI should never be allowed to make life or death decisions, as it lacks emotion, empathy and ethical judgement; however, it could be used to support humans with decision making by analysing and presenting the data and providing different insights/perspectives. Discussion followed about corporate responsibility if AI makes mistakes, especially in critical areas like healthcare.
• It was felt that AI will impact on employment and jobs, with some jobs becoming redundant whilst new jobs are created, and concerns that this will widen existing inequalities in society. Policies should therefore focus on upskilling and protecting the workforce.

Conclusion

The debate was a really rich discussion between the societies and AI itself along with good audience participation. Congratulations go to the IEEE City Robotics Society who were voted the winners of the debate and won £50 from the SU for their society.

Thanks go to the Debating Society for hosting the event and to all the students that took part. Feedback from societies and audience was really positive and there was definitely an appetite to run another inter-society debate. If you have an idea for a topic for a future discussion, then contact Christabel Carter.