Artificial Intelligence (AI) has made huge strides recently, with the rise of AI Chatbots such as ChatGPT, Google Bard, and Microsoft Bing becoming commonplace. Prominent tech figures, such as Bill Gates, have claimed that this is the start of a new epoch for humanity, the age of AI, and could be as seismic a shift as the industrial revolution.
Within the context of education, AI is both exciting and worrying in equal measure. Therefore, it is worth examining some of the potential opportunities and issues that might be presented to us. This blog post will hopefully serve as an introductory look at generative AI and how we should seek to approach it.
What is Generative AI?
Generative AI encompasses AI driven tools that can generate content which could take the form of text, images, video, audio, code or other forms.
One of the most common tools using Generative AI are chatbots, these are AI driven bots that users can chat to by giving them questions, the bot will then provide a response and the user can then ask follow up questions which the bot can respond to. So you could ask the bot to tell you everything it knows about the city of Brighton and Hove, or to write you a song about the planet Jupiter from the perspective of the Moon, or give you a thousand word essay on the work of Keats, and it would produce a response. View a recording of a bot responding to a query.
These chatbots are essentially language models that are trained on a huge corpus of information, often taken from the internet or another source, when a bot creates a response, it doesn’t really understand the question but rather the is using its corpus to locate the words with the greatest possibility of being related to the user’s question and being related to the prior words in its response. In the image below you can see the probability the bot has given to the most likely words and it will use this to generate its response.
The main thing to keep in mind is that these bots are not truly sentient or alive, bots don’t understand what they are saying but rather present a very convincing illusion.
How should we approach Generative AI?
AI will be something we will have to deal with as an institution. The University of Sussex is taking the approach that rather than attempting to ignore or ban AI we should embrace it and use it to enhance teaching and assessment. There are a few good reasons why we should do this.
The first is that there is no effective way to ban or ignore AI. Students are aware of AI tools and will start to use them regardless of what we do. It is also incredibly hard and potentially impossible for a tool to identify if a piece of work has been generated by AI and any tools that do emerge are quickly being made irrelevant due to the speed at which AI tools are improving. Engaging in a cat and mouse game of AI tools and AI detectors would simply be a massive waste of time and effort.
A better approach is to educate students on what constitutes appropriate use of AI and what would constitute misconduct, so they know when and how it’s appropriate to use such tools. If we engage in transparent conversations with students then the vast majority are likely to be honest and transparent with us in return. Student want their assessments and degrees to have authenticity and integrity and are as concerned as teaching staff that this integrity is maintained.
By embracing AI we also level the playing field by showing all students how to use AI appropriately rather than creating a digital divide where more tech savvy students will understand how to use such tools and other students won’t, putting the second group at a potential disadvantage. If we embrace and educate around AI as a digital skill then we can close this divide.
We also need to consider that AI will be everywhere in the next few years and students will need to know how to use it both within their careers and general life much in the same vein that using the internet is now a required skill for most aspects of life. We would be doing students a disservice if we didn’t prepare them for this. We need to equip students with the relevant digital skills to use AI in an ethical, critical and practical manner, just as they would have to use any other tool.
Educational Enhancement now has a page on our website Artificial Intelligence in teaching and assessment that includes a variety of information and resources, this will be continuously updated to keep abreast of new developments in what is a very fast moving technology. Also look out for further blog posts in this series which will further examine the opportunities and challenges of Generative AI within education.
Please also reach out to our team if you have any questions or would like help using AI within your teaching and learning, contact us at TEL@sussex.ac.uk
Many thanks for your post, which I read with interest. I agree with your stance that our institution should focus on the positives which AI can bring. In terms of writing skills, I have seen students use this technology to compare their own practice answers to exams to a model produced by a chatbot. It does take a level of sophistication to use the technology intelligently, and we should be providing assistance in developing those skills.
I am though. less certain about the level of transparency you envisage. It is certainly true that the overwhelming majority students (or indeed anyone), do not set out to write dishonestly. However, the pressure on them to get a ‘good’ degree – a pressure borne of the financial costs incurred in studying – would drive the most diligent of students to bend the rules in times of stress. And we do live in times of stress, with the cost of living driving students to seek more gainful employment than before, difficulties paying transport costs to attend classes and so on, leading to a less rich learning experience. It is not the techology itself that scares me, but the little time available for students to learn how best to use it.
Glad you enjoyed the post and thanks for your thoughts, its a valid concern around where to find the time to teach student how best to use these skills and is unfortunately another onerous thing to place on students who have more then enough on their plate already, but it’s something that students will encounter regardless so we need to try and address it as best we can, hopefully the direction we’d like to move to is whole curriculum approaches where we can embed and build these AI digital skills throughout the curriculum and make space for it but are aware there’s no quick or easy fix to these sort of issues and it will always be a compromise to see how best we can equip students within the limited time available.