The future of GenAI in Higher Education with Peter Hartley and Sue Beckingham

So this is Sue Beckheim. I'm an Associate Professor and National Teaching Fellow at Sheffield Harlem University. I am Peter Hartley. I am an educational consultant and visiting professor at Edge Hill University, and you are listening to the Academy's Developing Practice Podcast. Hello and welcome. In this episode we chat with Sue Beckheim and Peter Hartley about the future of Generative AI in higher education. And rest assured, we did have Peter with us in human form and his introduction was merely an experiment conducted with chat GPT. We hope you enjoy. So and Peter were really pleased to be speaking with you today and we're delighted to have you on the podcast to talk a little bit about Generative AI and its use in higher education. But before we get started and we do this with all guests that we have on the podcast, could you both give us a little bit about your background and how you have arrived in the positions that you found yourself in today? Okay, well I suppose the short answers are a very long interest in tech, computer technology. Just thinking back to when I first started playing with computers. And when speech recognition meant that you had to speak in a very monotone, even voice and you had to leave a space between each word and we wondered why it didn't catch on at that point. But you can pair that with what you can do now that is. So getting students to use tech and see how it works has been a long interest of mine. Met too many some years ago and we started collaborating and so when chat GP, chat GPT appeared, it was this is a change, this is a fundamental change and we need to get on board with it. And so we then and we just finished doing a a successful group work book, a little handbook for students and we were starting to think about technology and then we started discussing chat GPT and then we thought we'd better get this into Cedar so we did conference workshops for them and we decided after the first one we had a shelf life about half an hour because then Microsoft and Google suddenly came along and announced that they were going to put in all their products. And so since then that's been a roller coaster because we've just been keeping trying to keep abreast of it. We've done sessions for conferences and events. Last week we're at Cedar 2024 conference and you'll find our slides online in a month's time we're at Edge Hill and again, the hourclists of slides will change by then and also we've got a little bit of a sewer tell us, we'll tell you she's got a we've got a co-edited book coming out through Cedar so it's been a bit of a whirlwind and I'm in my third career now because I've been teaching academic then I moved into education development and now I'm working part-time semi-retired but I don't know how long that's gonna last because generative AI seems to be expanding. So yeah very much for what Peter says but again myself been interested in technology and hands learning. Specialised a lot really around social media for learning but really like to experiment and try try different approaches to make teaching interactive and get the students involved with different things and play really and you know that that's kind of how it's arrived with the generative AI because you know we'll say right from the start we're not experts but we like to sort of look at how these different tools might be useful to us or not as a case case maybe and and then you know sort of discussing the implications of using these tools which we'll get on to later in the the discussion but yes we've got a book coming out called 'Using Genose of AI Effectively in Higher Education' a sustainable neticle practices for learning teaching and assessment which we've co-edited with Professor Jenny Lawrence and Stephen Powell so that's a route led to Cedar focus group so that was a whirlwind putting that together but there's some really excellent chapters with people across the sector both nationally and internationally so yeah it's been great to learn from them and I think that's one of our big ethos is really it's having a conversation constantly with other people because we're all learning together and that's that's peers as well as our own students who you know some of them are ahead of the game completely. Well I bet it's really hard to be ahead of the game how are you both coping with the pace of the change in generative AI because as soon as I learn something it's it can be obsolete well that's what the first thing we wanted to say was can we come back next week because you haven't quite got on top of it yet and there's some well there was a conversation I've read online this morning actually between two supposed experts on AI arguing about how it works and I think well if if the experts people who've been working in it and looking at it for years one of whom was a very distinguished scientist who did some of the technological technological developments early on is hey you got to agree what hope is there for us there's a lot of stuff comes out on youtube on a daily basis so I think one of the one of the perhaps one of the tips for anybody who wants to try and catch up on this is find your gurus because there are a number of folk who are producing some very very good and laughily not too long youtube pronouncements which actually does help us keep up to date but the as you said the the rate of change is astonishing this first question so it sounds a bit silly in light of what you've just said because it's what we want to know is what is the current landscape for the use of gennämily in higher education and by currently it sounds like we might mean within the space of this actual conversation now but I think it would be helpful for the people listening to the podcast to find out what you think the current state of HR and gennämily is and then at least that will give them a sense of where we're going in future well the the short short answer is that it's fluid and we're currently riding the wave you know it's forever changing and developing so the currently four leaders um three that have really kind of come up and announce what they're doing and their developments which is open AI, Microsoft and Google and kind of waiting in the background is is Apple and and they've definitely got things up the sleeves but we think that they're biting the time and you might have seen some of the news releases where there's been certain little things that have not gone quite a plan that's Google have done and well all all three are the big big ones so I think Apple are waiting on so you know they're there in competition with each other you know there is a lot of money invested in this and there's definitely a lot to be made from it um so that's the current landscape and I guess we've lost count of how many different other apps are out there and some of the ones that Peter and I have found and followed have suddenly disappeared and I think what's going to happen over time is that the big four will kind of absorb some of these as they tend to do with new technologies so it'd be interesting to see where the independence can get bought out and yeah if they've developed things independently whether they're going to make some money out of it. I think that's one of the one of the issues over the next few years is given the amount of money that's been rolled into all these startups and also investments by Microsoft and Google etc is what is the kind of long-term economic picture for this um you know all the folk who are saying we should be using effectively generative AI more or less all the time 24 hours a day but what's the cost of that because there's some serious environmental impacts there's some serious questions about who can afford it on our sessions of the last year so one of the questions we've always raised is that will digital divide you know if the really good version of chat GPT is $20 or 20 pound a month then who's going to be able to afford that who's going to which institution is going to be able to afford the licenses for it now it may well be that that picture changed day before yesterday because OpenAI said that our our best new model which is chat GPT for O and O stands for Omni and we can talk about that because that's a really significant change if that is now coming free to everybody then what does that mean is that suddenly rewritten the rewritten the book um because if that's now freely available everybody can access it without too much investment then that's a significant change a significant step change I think another significant step change is the fact that I've got chat GPT on my mobile and you will have it as well um there's also implications for the others the competitors because of the you know they're charging or have been charging so there's Microsoft co-pilot you know if there's now going to be this free version it's actually a good version, then they have to do the same. And that with a lot of software, as you're probably aware, there's always the tiered, you know, they'll give you so much free and then they kind of dangle the character and actually you can have this extra stuff if you pay this premium. But you know, if it's enough that people can use it constructively safely and ethically, then that's that's good. Whether the larger business is sort of the corporate versions, just just as you have enterprise versions for, you know, communication with within organizations that they're prepared to pay for them, but we'll have to wait and see. And just to suggest another trend that happened, which I became aware of in the last couple of days, one of the things that we've noticed over the years is the different partnerships coming up. So open A have now partnered with Microsoft and Microsoft invested a lot in open AI because some of these large language models that are the basis of all this technology are incredibly expensive to build and need incredibly expensive and an incredible amount of very expensive hardware to do. Well, one of the things that came out yesterday from the Microsoft announcement was they've partnered with Khan Academy. And I don't know whether you could you've come across that, which is this kind of mass tutor. Well, they had salt on there to talk about the part, the impending partnership. And one of the things he said was that they had their ambition is to put an electronic personal assistant to give that freed every teacher in the USA. And that's a bit of a change, isn't it? If this sudden technology is there, available free. And if you've if you've watched any of the videos, I know there's some criticism of Khan Academy, which I'm trying to explore at the moment. But if you watch the way that they do it, they don't just deliver the answer to the to the question or the problem, they talk you through it step by step. Now, I just wish I had had that when my kid was doing a level mass because we never quite communicated because his understanding of mass, my understanding of mass was somewhat disparate. We never quite better the middle. So something like that, which talked you through acted as a coach, not as a deliverer of the answer. And if that's free to every teacher, well, what are the implications of that? I think the next thing to consider that we're very much aware of, we've done a number of talks as I say over the last year, 18 months, is that not everybody's done the same page. So there are lots of colleagues who have not even tried chat GPT. There's the fear of the unknown, you know, is it is it is it dangerous? And one we've done a very quick whole beginning of sessions. In the main 50% of participants had never tried it, which is quite astonishing really. And that could be the same with our students. Although we do know a lot of them are using it. And we're simply don't know. They're using the free free versions for obvious reasons for accountability. But there's there was an excellent study will buy Louise Drum from Edinburgh, Napier. In fact, she's on to a third version of it actually where they've done an anonymous collection of information from students at their university to find out how they're using it. And they've come back to say yes, they're using it to support their thinking and writing. They are aware that there are ethical issues of using it. The beginning to navigate the affordances. They've used it as a dialogue partner. So you know, sort of question and answer backwards. They found that it provides timely support. So out of hours, you know, if they can't contact the tutor, that's that's given them some help to a certain degree. But they are aware that it has implications and what they were asking or some in the response to it was the need to be taught how to use it well and how to use it to prepare for future work that they're going to be engaging in when they go out into the wide world once they graduate. So they're eager to learn how to use these technologies in positive and ethical ways. I just I just happened to be I just happened to listen to a podcast last night from America where they were talking about reactions to chat to you tell. And again, they've done surveys and it's interesting that the international comparisons you can find on this. And they found that percentage wise, twice as many students are using it as opposed to faculty staff. So they're probably in the same position that we are. And at the CD conference last week, we were we were doing a joint presentation with colleagues from from Banger and Abris with who've done a survey on their students. And one really perhaps frightening result that came out of that is that most of the students that they'd surveyed didn't really know what their institutional policy was about is the use of this. And you know, that's that's quite frightening because if people are worried about am I going to be caught by the plagiarism detector which doesn't work anyway and then we're going to be falsely accused, what do I do? And we know that this technology can really help some students, particularly students of the second language because they can they can prove free, they can check, etc. But unfortunately that could well be supporting a style of writing which actually makes some plagiarism detectors wave the red flag. And it's how you communicate that policy so you could have a very well crafted policy that you know people have worked some time on, you know, and it's hidden somewhere on the internet, you know, you know, it's like I, this thing's there, it's all very useful, but you know, do people know where to find it? And recently, co-wrote a booklet on assessment with Advanced HG with some some colleagues. And part of that was the assessment brief and at the University of Chester, they've now got a statement on their assessment briefs, it's a standard that says, "Our students allowed or not allowed to use generative AI and how." So there's these kind of things that makes it sort of in the eyes of the students and sort of really brings it to the forefront to let them know what's possible and not possible. I was going to say the other thing to to consider is how universities have responded or academics within universities, leaders within universities, it could be a mixed mixture of those. And there's a very good blog written by Philip Hardman, and you could follow up in the notes. And she says that there's been three three responses to generative AI and that's Team Vito, who's absolutely no way, completely, we're banning it, you know, it's not going to happen as if it isn't going to happen underneath the radar. There's Team Bypass where, you know, let's get back to pen and paper, exams, exams, exams, that's the only way to solve the issues. And then there's the more positive team embrace. So Team Vito, bypass and embrace and yeah, is there. We think that the dialogue around that has changed in terms of the reaction that we've been getting at webinars and events is that there are much more. And actually, that was reflected in that podcast I listened to from America, where it's faculty, as they would call it, are much more interested in trying to use this technology for positive purposes. So Team Vito and Team Bypass are probably. That's the people that are attending the session. Oh, that's true, yes, that's right. Absolutely. Actually, on the ground, I think we've got a slightly different picture, you know, and then conversation goes, dynamic integrity and all that sort of thing. Would you both be advocating that educators catch up? Because you've mentioned some stats there around more students are using it than educators at the moment. Would you be advocating that they dive in and start to have that sort of play time with chat GBT or other tools? Yes, you've also put the short answer to that. And also, I think actually, I listened to the previous podcast in your series from Charlie. Charlie race, who I know from for a previously characters with Cedar. And we are speaking very much from the same sort of, the same text in that he was talking about getting involved and participating in the student. We're saying exactly the same, because nobody is an expert on this, because it aids us so new, it's changing all the time. We don't really know it's how what its capacity is going to be in six months time or a year's time. So we need to actually approach it as a problem to be jointly solved. That's the way I would think about it. And also, because it has very different implications and for different disciplines. So we're looking forward to a session that we're delivering tomorrow to music educators in Ireland. and if you've been following the way that music generation goes, where you generate a music clip from a quote, and that started out as 10 seconds, 20 seconds, well you can now generate a full three minute June, and it's can sound very convincing. And there are about three or four, again similar to the way we've got with Chanty PT, you've got three or four big companies competing to try and produce the most realistic, the most authentic sounding from a text prompt. And I've tried it and it's not quite there yet, it makes some very funny choices. But again, I haven't experimented with very detailed prompts yet, but that's this weekend's fun. But it can only get better, can't it? I look at things like stable diffusion and mid-journey, and I was using those and like just in the six months that I first started using them to now, the things that they produce all wildly different and so much better. So the rate of improvement is massive, it's huge. So I guess this will be the same thing when it's generated music. I have heard of some tracks actually that then turned out to be AI generated tracks in the style of an artist, and a big Beatles fan. So I hear a lot of new Beatles music, it just blows me away. And that, of course, is this last Ted Ticket of Conversation is packed full of copyright implications. And what can you ask the software to produce "Alar, Alar Beatles" or what have you or. So for example, I don't know whether you saw the newspaper headlines over the last 24 hours about Scarlett Johansson and Open AI, the fact that they appear to approach the stories, a bit muddled of the pedigover who you read and what you read. But the truth is, they said no, and they've done it anyway. So copyright is really interesting, it really is together there. And it's early, it's like the visual and the music and what have you that kind of highlight those and bring them to the fore. So every subject is going to have slightly different issues, but every subject here is going to confront them in some shape or another. So the other industry that's going to profit is the legal. There's a lot of soup starts to come in for various different things. But there's another point actually coming out of what we said a few minutes ago, which is that every university has probably now got some piece of paper which says that these are our principles for using AI. But when you look at them and then you think, well, how is a student going to apply that in Prana, remember staff, how are they going to apply that in practice? Then you have a rather different question, you know, so for example, I've just produced my essay, oh, I use Grammily. But Grammily is now using Jenny, I am about that grand if you have the most sophisticated version, have I now just committed this in. So I think there's all sorts of fussy areas that we're going to have to encounter, which is why staff must work with students. Yeah, so imagining that colleagues are in team embrace because they're the ones who are going to be using this. How can they design curriculum assessment to respond to both the challenges and benefits of AI in H. We were thinking about this having the other day actually and we thought that first of all, start exploring it capability with students. How are they using it? What have they used before? What have they found out about it? I think that's a sort of conversation we should start having the students on day one of year one because already schools and colleges are well as there are certain experts trotting around schools and colleges now advocating very strongly the use of Jenny, I. Some schools have gone in for some of these packages that incorporate Jenny, I kind of under the hoods to help. So students will be familiar and students are familiar with with chat, GPT almost by default, I suspect now. So it's a conversation we should start having with students on day one, it's a conversation we should be having with them in terms of well, what do you think you should be using this for? If you think about your possible career in future, how's it being used there? So I mean, there's a question you can start asking from day one and encourage you that's kind of safe and ethical because nobody knows its ultimate potential. Nobody knows in any kind of detail what the impact will be on the world of work. We already know companies that have lay people off because they can do, they can automate particular parts of say the customer service chain. The greater or lesser advantage, so there's going to be impact on the world of work so that start the conversation with students on day one would be one of our recommendations. Do you have anything to add to that? Yeah, I was going to introduce you to the professor, August a cause is created this little framework called pair which is problem AI interaction and reflection. So the problem is it's getting students to either identify analyze define problems before they start using the tool to think about what is it that they want you to find out. And then they can look at the AI tools and explore how it can be used and of course there's a variety of different tools. You could have students going off in different groups using different software and then coming back and discussing which was useful. So you know the interaction with them, you know comes back to the prompting you know the there's the bad in bad out or rubbish in rubbish out, you know, it's the questions that they ask that critical thinking of how you might add on to that which is quite useful and chat GPT because you can build on a question of you keep adding to it and asking it to present it in slightly different ways. So that that's helpful and then the most important bit at the end is reflections for this willingness to sort of look at what we found what other sources can we find that a little more legitimate you know the literature. How do they align up and you know and having some examples sometimes of where it's completely made something up or it's you know it out of date or whatever the situation is in your particular discipline top topic so that again that can be discussed that you know we can't take this at face value you know generative AI is bringing together. The good the bad the ugly that the internet or you know the flip side is that you know the there are useful things as well so I think getting them to think about that and actually we want students to have good problem solving skills and critical skills so you know making that part of the activity supported by the tutor. You know and then working along because there's going to be some surprising outcomes I'm sure with with information that they'll think actually that's that's helpful or I know my topic and that's completely you know out of kilter. Yeah one of the first things I learned in GCSE IT was garbage and garbage out that yeah and one of the things that what fascinates me about generative AI is that. Garbage in garbage out still still applies you can through through that sort of transactional approach with chat GPT you can have garbage in and eventually it can change path and you can get good out I think that's the most fascinating one of the most fascinating things for me when I use it is that I can start off in the wrong place but get to the right place. What also I can start from the wrong place and go down such a long winded rabbit hole I can never get out of it and I can imagine a student might be in the same boat as well so they might you know they might use chat GPT or another tool to to find answers or to summarize maybe content and actually it's not given them exactly what they think they should be getting exactly I was compared to Wikipedia you know I was just trying to get that out of the way I was. You know we still have conversations with students the Wikipedia is a good start in place but you know you've got to be care thoughts in there we don't know who's written it and their expertise behind it you know with Wikipedia obviously you've got potentially some some references to follow follow up. And you know chat GPT in the like is you know a bit like Wikipedia on speed you know just does it fast but doesn't necessarily give you the right references as but having said that things are improving so. And remember that if you think about the announcement that came out of open air this last week we can now talk to our GPT we have to type in the question and we can interrupt it. So if we ask a question or we ask it to talk about X and it goes off on the wrong track and say hang on and it will stop. So the the notes it's become much more you use the transaction or that's one way that it's going to develop dramatically over the next however many months however in weeks however many days because this idea of having it as a conversational a conversational buddy your conversational assistant is a very different notion and just typing in your question and just waiting it for to finish and then I didn't like that right and let's start again. So the whole kind of interactions that we have I mean the open AI presentation a few days ago is really it's worth having a look at some of the examples that came out of that because unfortunately they have used a very cheesy personal assistant who is full of awesome and have a nice day. Of course and if you want to and if you want to be supported by someone who has that token of voice and that style all the time and you're very welcome but I personally I would like somebody to talk to who's a bit more straightforward and a bit less gushing. So there's some interesting thoughts going there's interesting choices that you can make going on there but this idea of the conversation with the with the AI rather than just the interrogation the question and answer. I think students are going to respond to that but provided they get the right personal assistant to play with. Yeah that's what I think that could be really powerful. Yeah I think that's I presume that people if not now then quite shortly people will be able to choose the voice that they have so they can have their teacher voice so that it becomes perhaps an more educational experience for them. So yeah that'll be interesting so there will be this whole range of sort of you know for so normal so the Gen AI can have that the person can choose to engage with in that particular context. Well that's already that's already an existence to some extent there's a software called character dot AI where you can choose the character you wish to talk to. So if you want to have a quick chat with Aristotle there you go and that what it comes back with is quite an interesting rendition of that person. Yeah you've also got the kind of personal the kind of personal well-being type AIs and the whole range of them with various degrees of friendliness and again they can be very convincing in this notion of so until three recently they've been kind of text based but if you can now talk to them it's a whole different ball game. Oh yeah absolutely. Yeah it's it's interesting that sort of coach in approach that AI is taking them there's a tool in LinkedIn learning which as the they've developed a little built-in coach so it because it's connected to your LinkedIn account it knows what you do for a living and if you put there some sort of goals in there it sort of knows what you're after when you're in LinkedIn learning so it's you have a conversation with their AI powered coach as they call and then that directs you to LinkedIn learning content. I thought that was really interesting now I guess the next step from that is actually rather than it directs you to LinkedIn learning content it just coaches you through the content itself yeah and that's I think that's fascinating stuff I think we and if how what I hadn't seen the open-eye AI's statement on that so how far are we from having that in terms of access for all so is that going to be a paid thing and then it sort of trickles down like the rest of the tools well that's the dead that's a big question really I mean I know the announcement they made was that this new version of chat GPT will be freely available now we have that there will be limits to how much you can use it but you will need to explore over the next few weeks and months just how free that is how far it covers what devices it covers like mobile but I know I think there's another interesting really thing that came out of open AI which is that since they introduced this this you know the level behind the paywall where you had to pay your £20 a month one of the things they did there was they allowed people to construct their own personal GPTs and there's I think several million of them hiding behind the chat GPT paywall but if that's now accessible to everyone so the notion that you can construct your own little personal chat GPT based around perhaps a particular area or a particular topic and you can feed it information on that topic as well now Microsoft have announced a very similar idea that you'll be able to produce your own read GPTs using Microsoft so that could be the thing that we ask the tutors can capitalize on over the next few years the notion of having your own little thing so I'd add to that is whether universities for example could have their own so it was kind of a bit like how we use blackboard you've got that walled garden of so you know it's safe that that information's not going to be replicated and used all across the globe it's just your and you whether you own it or not is a move point but you know it's at least private semi private to the institution because you know there is a concern about how people are using it for marking and put in in student work and all of these things there's the the ethical considerations of factual areas all that's going and how is that then feeding into the overall information that it's then replicating and using and other things which um yeah doesn't make it verified information so can we come back next week because we might need to revise something of what we've said. One of the things that um some people are asking me about you know organizational development and what that looks like in the future with Gen. T. V. A. I and my one of the answers I think I come up with was that I wasn't I'm not concerned about it in terms of my role like I'm embracing I'm in the embrace um team um but I I come around to the conclusion that AI won't ever really afford to understand the culture of the organization that I'm working in but actually what you've just said with sort of maybe university-owned AI's that's not quite correct is it that no I think that's a really interesting point because I mean if you look at the the sort of road map that Microsoft is setting out with the with their co-pilot and and what have you effectively they will the machine will have lots more information about the organization than any one person in the organization yeah simply because it will have collected it all and be able to search it and it will have a mechanism in other words the large language model the chat GPT type interface you'll have a mechanism for searching through that um so the other things the productivity affordances so one example is um with the paid paid versions you can have meeting notes also must be taken with a piece of software you know and there's implications there so yeah it's going to save somebody you know from professional services taking those notes because actually the the AI can do that for you and bring it together but um Michael Web who um is an expert in AI and works works for Gisk he's just recently said that you know we need to think about some etiquette rules regarding that and it's it's quite simple and obvious but it needs to be sort of spoken about and um yeah so don't invite a service if you're not the organisers of the meeting because I've heard of situations where people have suddenly realized that some things in the background and nobody's been given um that information or or agreed to it the second is the um use the tool that your organisations looked at to make sure that you know that the data's um safe and all that sort of thing and and that at the beginning of the meeting everybody should be told we're going to use this tool does anybody have any objections in the subjections then we don't use it for that meeting so it's three simple things to make sure that it's um used ethically and safely um but you know these are the kind of things that I think will emerge it's like when we first started to use the internet you know and the etiquette rules came came out we really need some some rules here which when you read them this email is but actually it stops those things happening because people just I didn't think you know yeah one of the things that Microsoft announced the other day was it was team co-pilot so they're trying to make give teams the kind of co-pilot type of intelligence um and they were kind of reshuffling their their software provision to make that more more likely and also more helpful so the notion of you know in the the the meeting that automatically generates its own minutes yeah keeps that you can interrogate in the meeting so you can say, have we discussed this before? Oh yes, you discussed it on September 14, 2022 and this is what you concluded. So useful, useful stuff like that. But I mean, that notion of the machine knows everything about you in a sense. It has all the information there. What's it gonna use it for? How are we gonna use it? - Yeah, I mean, in terms of my role, that's great 'cause I can interrogate that in terms of what, you know, and then when I'm talking about the culture of the organization, our performance and where we're going, that's a really useful tool for me to interrogate, whereas, yeah, previously I thought I would never be able to get that. - No, no. - And can I just, you just remind me, can I pick you up on the thing you said a few minutes? - Sure. - Where you said we must join team engage. What we would say you must join team critically engage. Because I think, you know, this technology can produce complete rubbish at times. - Yeah. - And completely misleading stuff. And also it's still getting better every day. So we've got some lovely illustrations that we use in the webinars of illustrations where the number of legs under the table don't match the number of bodies in around the table. So there's all sorts of peculiarities like that. - Yeah, because I think you've both touched on the issue of work. And, you know, both in terms, perhaps, of AI replacing people. I mean, there's been a lot of stuff in the creative industry. It's around people using their jobs because, um, Gen AI can create copy that is passable, um, you know, as being written by a human. But also, you know, just now we've had this discussion about how Gen AI can create meeting minutes and things like that. So, you know, in terms of sort of students, you know, how can we develop our students' employability skills in relation to the use of Gen AI? - Well, we already, they're already using it to help them write legal letters and CVs, but we know that, but we also get feedback from employers that, you know, critical at that use because it's quite apparent in the style of writing that they're using that they've not written it. So, for example, typical CV, you need to give examples of skills that are on the job application. You need to give authentic examples of where they've actually used those skills in a given situation, which, you know, if don't make sense or even if they look passable on a CV and they get to interview, they can't answer the questions. Because, you know, they aren't thought through, they, you know, the employees are actually going to look at this. So, you know, I think we do have a duty of care, if you like, to help our students understand how they use these tools and use them for good things. But they've also got to be savvy enough to check the information that's that's coming out of it. And, you know, this, we've talked earlier about Grammily in there. And there's certainly lots, you know, we have many students that have got learning contracts and, you know, sort of writing is not their strength point. So, yes, it can help them. But it's got to be done in a way that they can then be sure that their authentic voice is there in it. So, I think that's important. And as Peter said, and we've said so before, you know, critical awareness of how they're using these tools. - So, looking ahead, what do you think will be the landscape for the use of generative AI and high education? Maybe in 12 months, I mean, it's really hard to say it's, because it changes so fast. Maybe I've actually assorted that and go to six months. What do you think the future is going to look like? - My answer is, what would I like to see? - Yeah. - Because we don't actually have a crystal ball, so we can't actually be sure. But I'd like to see that it's inclusive and accessible for all. So that, you know, everybody's got a decent tool and appreciative that people want to make money out of this is always going to be the tiered. But there's a tool just as we can use Google to find information that, you know, people can have access to that. And secondly, it's less damaging to the environment and where they get the information is more ethical, because there's obviously concerns there. - I would endorse both of those wholeheartedly and maybe add a third thing which we would like to see students and staff working collaboratively to explore this technology in terms of what it can do and what it can't do and what it should do. And with that, getting the potential to do that is closer this week than it was last week because of the way things are developing. And next month, who knows when we decide what our iPhone will be able to do in the fourth coming six months? - I think as a prediction, it's going to be an accepted useful past assistant. Hopefully, you know, it's going to be improving month on month and it'll be everywhere. So you'll be able to get to on your mobile, your laptop, PC, your watch. Yeah, maybe you could tell me where my bus is, really. - In real time. - You can still, you can already do that soon. I have a little thing app on my phone. - So do I, but it's not real. The bus doesn't come when it's supposed to come. - Well, sometimes it depends on the company. Some of them, they disappear, which is certainly disturbing. As you imagine, the bus has literally gone off the road, but no, it's been interesting to say. - I think the other thing that might happen, and I'll be interesting to see where this goes, is that whole notion of programming stuff for yourself. You know, having a personal GPT. - Yeah. - When I got various, you know, go about years, handouts and stuff like that. And I thought, well, there's some good stuff in there. There's not quite there in the textbooks. So can I use that to develop my own little personal kind of store? - And we'll still be able to talk to the VLE, Google or Blackboard and say, do you know, I remember having that class on whatever talk to this is. Can you take me straight to that context? I can't quite find where it is. Or can you then create me from those slides, some revision questions, or, you know, there are all these things that kind of technology is, but it's not all joined up. So yeah, hopefully it will be more useful. - I think that'd be wonderful if you could say, you know, provide me with vision questions on, you know, X, Y content, that'd be fantastic. If you couldn't find it already. - Well, you know, you know, you know, already do that in general, but it's this notion of being able to home in on your, - Yeah, on your, yeah. - Yeah, it reminds me of, like, when I finished high school in the '90s, there was some in the off the internet was taking off. And there were some browsers that you had to pay to use. And some, I tried to be free. And it reminds me of that sort of era where, - Yeah, yeah. - All these things were happening. And eventually they, they converged into, you know, maybe the big four or big five. And it then became free access for everybody. You know, everybody just, it became the norm. You know, we just use browsers. I think it sounds like it's good. - So maybe I'll listen to, yeah. I used to have to pay for AOL. - Yes. - And then, you know, Google's free, but then you have to sort of accept that, you know, Google also knows what you're emailing about and actually can tell some things. - And another thing that came out to the announcement just last week was Google saying they're going to put AI into their internet search. So a Google search tomorrow is not gonna look the same as it did yet last week. - Yeah. And they're trying to, they're having to try and do that in order to keep hold of their income. But also, we'll see what, see where that goes. Because one of the assertions from that year ago was that, you know, all chat GPT would somehow replace Google. Well, that hasn't happened. So, you know, what's this next generation of Google gonna look like? Which will probably be here tomorrow. - Yes. Yeah. Yeah. When we talk tomorrow afternoon to discuss how that will be. Will we be able to use it? - Well, it's the only, this is the only area that I've taught in my career as an educator, where I've had to date stamp every slide. As by way of assurance to say, this is what we thought then. - Yeah. - What we think today or tomorrow might be different. It's like the American Senator who was once accused of fabrication and got rounded by saying, all my previous statements are inoperative. I don't think anybody believed him. - So, this podcast is called the Developing Practice Podcast. We like to finish each podcast with three or four take home points that the listeners could reflect on in terms of their own personal practice. So, Sue and Peter, if you could give us a few tips of thoughts to reflect on in relation to your practice and use some generally in higher education, what would they be? - We thought about this the other day. And they, and we thought the number one should be review your assessments. And if the machine could do an acceptable job on an assessment, then you need to redesign it. So that notion of looking through and saying, well, what can we do and what, what could the machine do? And if then you decide that, well, a machine could do this, it could produce the 500 bird summary of X or whatever it might be, then you need to redesign it. And we think that if you are redesigning assessments, you should be trying to do it to develop the students' subject expertise and their understanding of Gen AI simultaneously as one package. So for example, an example we discussed with some colleagues the other day was if you've had an assessment which is basically a sure test A on X, why not put the students, perhaps get them to work in 2 or 3's, get different large language models to respond, get them to look at those different responses side by side because they will be different. And ask them to critique and produce an improved version. And we think that's probably A more interesting and probably a better learning experience for students than just go away and write your own interview, do I say? Sue. And the third one which we've kind of alluded to is to the Fandert difficult topic or difficult to do and then to use chat GPT to create a self-test sheet. And use it as a study buddy or revision aid to get the students to practice with that. And then the fourth one was having their staff and student sunpits, so exploring things together. We think this is really important and it's partially the experimentation of the tools, but then it's discussing the issues which we've talked about earlier. So the important things are copyright, the impact on the environment, the digital divide and the cost because until there is more things free that's always changing. The ethical aspect so so where is the information being gathered from and in what ways. And finally data security which is really important. So be thinking about if you are going to put information into some of these tools, what are the implications of that. So yeah, there are our four tips to take away I think. And we can end with one absolute 100% confident prediction which is that last point about getting staff and students together to focus on this technology. We will still be saying that in the years time. Brilliant. Thanks Peter. We thanks for your time today. That's been really interesting.