3. The AI episode

Welcome to the H L N S C podcast from the teaching mentors. In this podcast, I'm talking to our digital learning mentor. Mel Newall all about AI and its use in education. From the point of view of students using it from staff, using it to really get an understanding of how we're handling this new technology.

I'm Naomi Anderson. I teach geography and business at Ludlow College in Shropshire, part of the HL NSC group of colleges. And I've had a little bit of experience with AI in the classroom and thought I would share those experiences here. So geography is a coursework, subject and business I teach on the BTEC course, which involves the writing of assignments. And my experience is that students found tools like chat G P T really quickly. Very soon after they were out, they were talking about them and having a go with them. I started to get a few bits of work that didn't look like student work that I would expect. And so we had discussions about referencing and we had discussions about appropriate use and my main sort of observation at that time is that they tend to get used by students that are facing panic. There's a deadline due and they are absolutely up against it and need to finish something quite long, quite quickly, and then cheat. And I think that's probably quite common. I think if we think back to students that have used paid for assignments from places like Student room and other sources, that's generally, it's panic, isn't it? And deadlines that cause it. So in my teaching practice, I've tried to reflect on this and look at, okay, how do we get to a point where, There is no panic and the work is being done in a timely fashion, and we are doing staged deadlines so that students don't get in that position. And I think that's where I'm taking my practice with regard to AI. In the short term, at least when a student has done this. You can ask the student questions about the writing and if they can't answer them and they probably didn't write it. The key thing. Is getting the student to produce something appropriately referenced that is their own work. And that's where I can, as an educator, support those students to come on. Let's make sure that you've used the correct referencing in the bibliography of this work, and you've cited what you've used and way you've used it and you've commented on it. I think that's the, the best way forward. And that's how we've dealt with it. And where a student is clearly trying to get away with malpractice. Then obviously you are heading down the disciplinary route and the malpractice route with whichever awarding body, and that gets quite serious. last week I heard about some students that have been thrown off a master's course, quite a number of them at a university because they'd all generated reports using AI and hadn't written or referenced them themselves. And that's at master's level. So this is going on through, all levels of education where Independent works required. I did attend an OCR training session a couple of weeks ago, which was O C R, my Geography Exam Board. They gave the most fantastic CPD session. It's recorded and available if you look it up with support from J C Q on guidance for teachers on how to work with AI. It's out there. We have to accept that it's a tool we can all use and how to work with it. And I thought that was really, really, Encouraging and encouraged healthy use of the tools that are out there. Google, Bard or Chat GPT, or any others that come along. Healthy use of them to use them as a resource to understand that it's no different to another website source, really. All it's doing is doing a very advanced search for you and coming up with something quite coherent. So in the sense of a geography, NEA, you could ask ChatGPT to answer a question for you and phrase it in a sense that it gives you a very geographical answer. You could use that as a contribution, as a resource in your geography coursework, but you would then comment on it, use it to decide whether it would oppose the data you've already collected, the primary data collection analysis you've done, or whether it would support that analysis, or whether it would argue against it and then comment on the quality of that resource that you've been given. So I think there are really positive ways to use it going forward. I've just presented to my first year group the first lesson on preparing for their coursework, and I've been very open about Chat, GPT, and I think I will run a session, a practical session in an IT room next week where we use it and we see what it delivers back, and I will try and show them that it's not always correct. I'll try and show them that it, it cannot replace their own writing. It has to be independent writing, and it's a, it's another source and resource. It is not a tool for production of, of their own text. I think that's my only sort of way forward really with that. I need to be sure that they understand that the data in ChatGPT only runs up to 2021. It's not that recent. So they really need to think intelligently about what it is they're hoping this tool can do for them. That will be my best best use of the resource that's out there rather than find it, sort of seeing it as a, a restrictive, negative unhelpful tool. Because if I have that mindset students are going to use it in the worst possible way. I'm going to end up with having to fill out malpractice paperwork and nobody wants that. So yeah, try and try and use it as a force for good. Use it positively and stay one step ahead. So in terms of how I identify whether students have used it or not in their assignments, I will know that by knowing the quality of each individual's writing, their skills, their level of terminology and language they use. I find if I read something that's being written by AI, it tends to have a few americanisms in it, American spellings. And if the students on the weaker side tends to be those that do this, they don't actually make the effort to edit or check that stuff. The more able students don't bother in the first place, cuz they're cracking on with their qualification. So, yeah, that, those would be the things I would look out for with a bit of over the shoulder reading of work. ChatGPT Zero is a great tool. You can copy and paste a chunk of text in and it will highlight it doesn't say for sure. It just gives you that there's a high probability that this was written by AI and not a human or I've put text in and it's gone. No, this was a hundred percent written by a human. So it's got quite a clever way of identifying it. But you can't know for sure and unless you have a confession from a student, you, you don't know, you can't prove it, but you can prove whether they know and have the knowledge on what they've just written or not.

I talked to my colleague, Mel Newell. about AI, about how it could be used for force for good. Here's our chat

Here with me today, I'm talking to Mel Newell, who works here at Herefordshire, Ludlow and North Shropshire College and is our Digital Mentor. Mel, tell us a bit about your role. Hi. Yeah, thanks for having me. Naomi? Yeah, so I work as a digital mentor. So my role is primarily to work with teachers to support them, to use their digital technology in the classroom and to improve their knowledge and confidence in using digital tools. So we're here to talk about AI today. What do you see as the key challenges AI presents for educators? I think the most obvious challenge is the authenticity of work that they receive from their students. So there's a lot of worry and concern, rightly within education that this type of technology, could be undetectable. And they don't really have a grasp of the, the real knowledge and understanding of what students are, are giving back to them. I mean, it's, it's been a long. Time in terms of what we've been used to. And you think, oh, it's plagiarism, so that's easy. That could be checked. We know you have a broad feeling of, as a teacher, you will be reading this research yourself so you all know if it comes back at you, you absolutely know that's come back from something I've read that, and you can go back and find it. And it may be that it's not been referenced properly, but these, this type of technologies is, is. Trained to be more human. So it will give a human response, a unique human response. So I think that's a huge challenge for educators, that they feel that maybe that the authenticity, the accuracy of qualifications and work being given may not be a true reflection of what's out there. It's interesting what you say about the uniqueness. Because Have I got this, right? If I put a question into ChatGPT and you put the same question, we'll get different answers. Yes. So that really does pose a challenge cuz it's really easy to spot where plagiarism happened, where they've all used the same source and it comes back at you three, four times. You're like, oh, hang on a minute. What have they all found? So this could be really different responses from different students. Absolutely. And I think depending on the institution as well I mean, most teachers will know how their students write, so they'll know their style. They know what level of work they tend to submit. So if someone becomes an expert overnight, you might be a bit suspicious. But if you've got anonymized marking, which a lot of universities will have, that's, that's an issue because they have no idea, there's no knowledge of that student, how they normally write. So that's a, that's a real opportunity for us in FE, isn't it? Because we know that the students that are presenting the work, we know what we're expecting. And if it's totally out of that kind of, yeah. Unexpectedly good. Yeah. Or unexpectedly bad. They reward. Yeah. Yeah. Yeah, that's interesting. Yeah. But then I think as you get new students as well as this technology is more readily available, you'll have a new intake of students who you won't know. No. And if they're using it from the get-go, then their standard of work is that kind of, so we have to do more in class baseline testing to know those students earlier, get pen on paper. And I think as well, you might get a feeling cuz you think, oh, they don't say anything in my classes. They don't contribute. And I think it puts more of the onus as well on more formative assessments so that you then see in that environment without a summative, without it being summative, you can test their, their awareness. I think, jumping forward to some opportunities as well. It, it's opportunities for different types of assessment. Yeah. And more, potentially more interactive forms of assessment, more discussion types, and just testing and questioning what students know. And what they understand rather than just, write down what you know on our market and I'll realize what you know and what you don't. Mm-hmm. Yeah, that's really interesting. So you mentioned there's some opportunities. Tell us about those different techniques people could be using to assess the, don't involve written assessment. So I think you could use some oral work presentations, group discussions, professional discussions without relying on quite heavily essay or report writing. But actually I think you could use this type of technology, the, the AI as part of the assessment. So you could set that challenge, a basic prompt is put in, and then the challenge is, With what's returned critically, analyze or evaluate and broaden that and show how you use it for yourself. Mm-hmm. Or how you would improve on that response. Use it. So you are teaching students how to use the technology because we can't be naive. This technology's there. It's here to stay. It's already being used in workplaces. They need to know how to use it. And to evaluate that information. It's no different than if you are teaching them how to use the internet as a source of research for their work. Not everything that comes back from the internet is relevant. If they're using, if they haven't used the right prompts, they might get American information that comes back and then they'll put that in their assignment. You think, no, that doesn't apply. That's not the legislation here. Yeah, just using it as a tool and. Finding out what type of questioning techniques to ask to get the right response that you want back. Yeah. And those sorts of things. But I think you could also perhaps use different more different types of assessment in, in terms of more formative, so, Teachers are giving the feedback back. So the need for the students to use this type of technology is reduced because they're getting information from their teacher about how to improve the work that way rather than this huge summative at a certain point in time where everything is just on that. Yeah. Which is where you get the panic and deadlines and that's where people resort to there's been lots of discussions around this naturally, and we can't just go, they, they can't use it. Let's just ban it. We, we have to be realistic and we have to give students the benefit that they're gonna be honest and they don't want to cheat. But as you said, sometimes life gets in the way and then that's what they're doing. They're going into these things thinking, oh This will do the work for me. Mm-hmm. Yeah. So it's, it's early on making sure they recognize that we know that the tool is there and it needs to be used as a force for good. Yes. And we need to change our assessment so we know our students well and make sure that, if, if it is gonna be used, it's used in a critically evaluative way as another source of information and not the source yeah, absolutely. And that, that students recognize that using something like this, Is no different than if they used a textbook or a website and didn't reference that and didn't say, this is where I've got it from. Yeah. And building on that, because with this, this type of technology, the information that it's trawling through to generate its response, that may be someone else's words anyway. Right. And it's not referencing itself. No, exactly right. So there are, I think I've just been looking, I think Microsoft have got Bing Chat, which will reference its sources. So it will say, this is where I've got this information from, so students could then get back. That's positive. That's a good way of finding sources actually. Absolutely, and I think sometimes maybe people may not be aware that what they're using is actually generative AI. It's, it's available, it's in a testing site, but you can still use it. If it's there and it asks you to sign up and that's what you do. That makes sense. Mel, you've been brilliant. That's so helpful. Thank you very much. Was there anything else you wanted to say on AI before we finish off? I think that my final words on it would be, don't panic. I think it's just another journey for education and we will come out the other side and, there's lots of positive practice already happening and we just need to think. Hang on. It's not all bad. Let's not throw the baby out in the bath water. It's just a different way of doing things. Brilliant. Thank you so much for your contributions. That's really helpful. Thank you.

Thank you so much for listening. I hope you found that useful. Over the coming weeks, we'll be podcasting and interviewing colleagues around the college, about a range of teaching techniques and theories and concepts in teaching and learning, and we look forward to sharing our findings with you. Please hit subscribe if you've liked what you're listening to and leave us a review. We'd be thrilled to hear from you.