SPEAKER_02 0:00

Just to start us off, would you be happy to share your current role and your area of work or the area that you research or study in?

SPEAKER_03 0:09

So my current role is I'm a research associate at Adelaide University. The area of my research is kind of broadly in the area of physical activity and health, but I have a particular focus in activity promotion in healthcare setting and mostly in hospital settings.

SPEAKER_02 0:28

Amazing. Thank you. So when you hear the word AI, what is the first image or feeling that comes to mind for you?

SPEAKER_03 0:37

First image, probably honestly, the either Chat GPT or Claude interface at the moment, as well as maybe a bunch of stuff I'm seeing on social media these days. So I guess those two sort of areas, one in my work life, which I might use like Claude for at the moment, especially. But also, I guess my personal life, I see it a bit more and just like when I'm looking at social media and see just more and more content on there that is clearly AI. Yeah, that's probably the two main things.

SPEAKER_02 1:09

Yeah. Do you mean like you can tell it's generated by AI, like videos or images or like chatbots replying to social media like through comments and things? Or people using it?

SPEAKER_03 1:23

I think more, I think for what I mostly springs to mind is when I see like videos and those sort of content that are for the most part clearly AI generated. But I'm also like, I think the more I'm starting to, I don't know, scrutinize content coming through on my feed, um, I think I'm starting to realize that, oh, hang on, I think I might actually, I'm picking up that things I didn't initially think were AI, that they might be. Um, and you know, often that's maybe some things, and I've had a conversation with friends about as well, like things that are images of like nature or animals or those sorts of things, whereas like something that looks really fantastic or like um weird or just unusual or comical, like something that has, I guess, a bit of an extreme response that you're like, oh wow, like it gets your attention, and then you're like, hang on, I don't think this is actually real.

SPEAKER_02 2:18

Yeah, some of it is like so like believable though. I remember seeing the animal ones, especially like seeing a dog do something like amazing, like jumping into a pool and doing something.

SPEAKER_01 2:28

I was like, Oh, that's incredible. And then the comments would be like, Oh, this is clearly AI, like a dog doesn't do that. And I'm like, Yeah, it is. Maybe I just wanted to think the best of it because I love cute animals doing cool things. Yeah.

SPEAKER_03 2:41

And then I like, I don't know, occasionally I'll like send it to a friend or re post something, and then I'm like, oh, that's embarrassing. I've just like I've fallen into the trap because I didn't realize that at first, and then I looked at it again and yeah.

SPEAKER_02 2:52

Yeah, but it's getting very good, isn't it? Like even with um like humans, like real people talking, and you know, you have to really look at their lips and see, is this person actually saying these things? Like, you know, if it's like a really important public figure, yeah, or like a political topic, you're like, did they really say that?

SPEAKER_03 3:11

Political stuff.

SPEAKER_02 3:13

Definitely. Um, so you kind of talked about that you use AI for work and you know, other things, but what's one everyday task you'll happily hand over to AI and one you'll never give up?

SPEAKER_03 3:25

Um, I think something that I find it especially useful for in my own work is um so something that I have to do quite a lot in my work is maybe respond to uh headings on documents. So like it might be an application, it might just be a form, it might be like things for emails, and there might be constraints around like word count or uh the extent that you might use technical language, that sort of thing, where there might be certain parameters that have to be met, but I already have the content elsewhere and I like to use AI to help me um rewrite words that I've already written essentially to meet the parameters of what I'm doing. Um, those sorts of tasks can often take a lot of time and feel quite tedious because it's just I already have the information and it just takes so much time to edit it. So I think those kind of um text editing things are really, really useful. Um, I do find I often still need to rework a lot of what the outputs are that it provides, but in terms of the actual task of producing that text, it's incredibly useful.

SPEAKER_02 4:35

Yeah, definitely. Yeah. And like those just those admin tasks. Yeah, that's like really time tasks. Yeah.

SPEAKER_03 4:41

And then I think as well understanding, sorry, um, understanding information that's outside of my area of specialty. Um, so yeah, so lately I've been having to fill out a lot of forms where I've had to respond to queries to do with maybe like bureaucratic stuff that's outside of what I'm more easily knowledgeable about. So kind of getting it to explain concepts in more plain text is really helpful. So I can feel a bit more confident in how I'm responding to things. Yeah.

SPEAKER_02 5:09

Yeah.

SPEAKER_03 5:09

Is that like responding to people, like through email and stuff or just like if they have questions or I think for at the moment it's a mix of responding to people as well as like filling out forms and um like project registrations or applying for things. Um, yeah, often and like with a lot of our work, there might be cross-institutional, cross-sector projects that we're working on. So the types of um systems that they might use within the healthcare system, for example, and the sorts of jargon that goes into a lot of admin there is a bit different to what I'm familiar with. So being able to like have something explain in more plain text what it's talking about is really a helpful thing.

SPEAKER_02 5:56

Yeah, a hundred percent. I agree. I'm definitely the same. Um, now this is not particularly about AI, but in general, what's the sustainability habit habit you've picked up recently, big or small, anything?

SPEAKER_03 6:11

Um, I drive a hybrid car. So I got that a few years ago now. Um, and it's definitely I saved a lot of money on petrol. It's a small car anyway, and because it's hybrid too, it's saved a lot of money on petrol. And um, yeah, it's just a nice feeling knowing that because I do use my car a lot. It's not as bad for the environment as my old car was. Um, and my husband's also we're about to get an electric car too, so that feels like a nice shift. Like, very lucky to be in the position to do that, but um, yeah, I think that's probably the biggest thing lately.

SPEAKER_02 6:46

Yeah, good on you, like reducing like your CO2 emissions and things like that. Yeah, oh that's great. Um so this is probably a bit more AI focused now again. Um, when you think about AI in your field, what's the first thing that excites you or worries you?

SPEAKER_03 7:05

Um excites me is that maybe the ability for it to take over the mundane tasks. Definitely that. Um I think maybe something that worries me is maybe uh, you know how there's like AI slop AI slop that is, you know, production of slop across different fields. I guess it's the possibility for that sort of thing to come into research where say there's like um academic publications or projects that are heavily uh created by AI, which maybe from a connection with reality point of view, it just doesn't have the same as like what a human might in terms of like how contextually relevant a project idea might be, how uh how good the writing is, how strong the ideas are, how robust the um the conduct of the research might be. I think, yeah, I think it's the possibility for that sort of thing to come into research that is maybe a bit more concerning.

SPEAKER_02 8:21

Yeah, definitely, because it's not really transparent then what's happened if AI is kind of generating what they think you want to hear. Yeah. Yeah, exactly. What they think is the best way to do something if you yeah.

SPEAKER_03 8:33

Exactly. Yeah, and like you hear of um maybe I don't know, I haven't heard it as much with research, but like hear about someone's like chat GPT thread, it learns the person. So it provides responses to I guess keep the person happy and keep them engaged. And so you don't know then if that sort of sort of flows onto work stuff where maybe it's providing things that are useful for that individual, but perhaps it's a bit disconnected from the broader field of research that that person's working in.

SPEAKER_02 9:07

Right. So it might not consider like diverse populations and other considerations for like cultural considerations, other things like that, do you think? I think so, yeah.

SPEAKER_03 9:18

It can almost become like its own little echo chamber.

SPEAKER_02 9:21

Yeah, that's a good point. And what about you do a lot of research in the hospital setting or in that setting? What do you think about that in terms of AI in uh, you know, those kind of hospital care settings, like what things are maybe exciting and what could be potentially boring about that?

SPEAKER_03 9:41

The things I feel like I um have heard a bit about from people who do work in the clinical sort of patient-facing um types of jobs is some things that can be really good, is helping with like clinical notes and like report writing and that sort of thing, because those are like notoriously time consuming, um just kind of like really rote sort of things. And also like when you're writing notes at say, I don't know, after five hours of patient time or at the end of the day, kind of remembering details can be, they can, you know, your brain can overlap sessions. Um, so AI helping out with that, you know, sometimes it will be a thing that sort of listens in and then summarizes notes during a session. I can definitely and having had like a background as a clinician, I feel like that would be amazing. Um, I've also heard about it helping with like assisting with diagnosis for a range of different things, whether it's like looking at medical imaging or blood tests, those sorts of things. Um, obviously that sort of stuff I don't speak from the perspective of someone who has had that scope of practice where they can diagnose, but I understand that like, you know, that will also still always require oversight of a real person clinician. Um, but I have heard there's a lot of potential for AI to help with those sorts of things and can help with accuracy as well. Yeah.

SPEAKER_02 11:16

Um, I saw some was that something recently about um diagnosing endometriosis through AI-assisted yeah, screening of the imaging.

SPEAKER_03 11:25

Yeah. Yeah, definitely. So like more accurate detection. I think some there's been some instances where it can help with um earlier detection of some conditions. So that could obviously be really good for like yeah, early intervention, which is always going to be better. Um, and even just like, you know, risk stratification, that sort of thing.

SPEAKER_02 11:48

Definitely like timely, getting like timely results potentially. Yeah, exactly.

SPEAKER_03 11:53

Not having to necessarily send things off like wait for it to come back and then upload reports. Like I think cutting out those, and you know, there's sometimes there we conditions that are really time critical as well. Um, so you know, where it can sort of speed up the the um the process as well as the accuracy in some cases is really exciting.

SPEAKER_02 12:14

Yeah, agree. Um, how do you explain AI to people who aren't in your area? Um, like family, colleagues, um, and what do they usually get wrong potentially or their misconceptions if you've you know?

SPEAKER_03 12:29

Oh that is a good question because I've never actually tried to do that.

SPEAKER_02 12:34

Um, most people do know what AI is, right? Like, I mean, generally because it's so mainstream now, but I suppose when I think of sometimes when I talk to people about AI, it's like, oh, is that just like all it's I mean for me, just experiences like is that augmented reality? Like, is that something that it's you know, in that term terms of intelligence? And I'm like, oh no, it's like a little bit different. It's like yeah, so I don't know what's do you have any experiences like that? I mean, you might not have.

SPEAKER_03 13:05

I don't think I have a personal experience that I can pull from, but I guess something that I see like memes and stuff about, or just like you know, hear the way people might talk about AI just with friends and stuff, like the whole thing of it being, you know, like oh, I'm always saying please and thank you to Chat GPT because I don't want it to come for me like when the world ends. And I know a lot of that sort of stuff is like it's all just silly fun and games, but like um I guess there is that sense of um personifying it when it's like it's not, it's like it's a computer code that learns how we talk and kind of mimics that in a sense, um, and that makes us as humans feel like it's a person in some regards. Um so I guess, yeah, I don't know, I guess as an as a misconception, like just that whole remembering that it's not a person.

SPEAKER_02 14:03

Yes, yeah, yeah.

SPEAKER_03 14:05

It's an artificial intelligence, yeah.

SPEAKER_02 14:08

Yeah, yeah, that's it. And um, I don't know if you've um had any conversations with friends and stuff about how they use AI or Chat GPT or anything. Is there anything that you've had like come across from having those conversations where you're like, oh, that's you shouldn't be using it for that, or or you know, I feel like I've heard of people using it for help with responding to like messages, um, like just with friends, like oh that gives me such an ick.

SPEAKER_03 14:36

Like I don't know, just the whole you using Chat GPT to facilitate your own social um connections. Some I don't know, something about that just makes me feel kind of gross. Like, you know, I have friends who might use it for help with all sorts of life stuff, whether it's work or study or like personal budgeting or like help finding sales on things, or like um what else? Like exercise plans, whatever. But something that just makes me like feel so icky is just the thought of people using it to help with interpersonal things, because in my opinion, it's like it doesn't have that sense of a social compass.

SPEAKER_02 15:25

Yeah, no, I completely agree. I actually caught one of my friends out because you know how obviously if you copy the text, it says, I suggest you write this or something at the the start, something like that, right? And so you immediately know that it's chat GT. And she responded to me to me with that, and I happened to just click into it straight away and I saw it and I was like, Did she just use Chat GT to respond to me? And then she she removed it, like deleted the message and resent it without that on the top.

SPEAKER_01 15:54

Yeah. So close to being like, girl, like what are you doing? Did you just use like delete it and cover your track? So yeah, and I was a little bit like I'm not sure how I feel about that. You can't just respond to me. Don't have the time.

SPEAKER_03 16:09

Admittedly, there's things that there's things like that I can forgive, where it's like, oh, you know, if you're like summarizing information for something we're doing on the weekend, like meet here and it's gonna cost this much, and like bring this or bring that. If it's something a bit more um impersonal, then it's just like, oh, you know, it is what it is. But if someone's having conflict with someone and they're unsure of how to have that conversation, or they are responding to like someone sharing exciting news, or they're in a situation ship, you know, I don't know, it's that sort of stuff that makes me where there's like an an element of um increased emotional like awareness required. Yeah, that's just what makes me feel ugly.

SPEAKER_02 16:56

Yeah, 100%. Do you think then I'm just thinking about in the context of education in general? And I know a lot of like teachers and you know, potentially like university lecturers and stuff would use AI to respond to inquiries of students or potentially like in you know, primary or secondary school, like um parents. Do you think then how do you feel about potentially using AI to respond to potentially like challenging situations and how to make that a bit more I don't know, personalized or more genuine? Because I agree, like it does seem like disingenuous to tackle really personal things like that through using AI. It's a tricky one, yeah.

SPEAKER_03 17:48

And it's I think it's true, it's it's different again when there's um I guess in the context of like education, for example, like where there is a layer of someone having a duty of care, yeah, over another person, like a teacher over a student, for example, um, where in a sense they are required to respond in a certain manner. And like I would imagine, I mean I'm not in education, but like I would imagine they have to deal with those sorts of things in a way that complies with some sort of policy. Um, and so I think in the sense that perhaps when say a teacher is responding to something like that, I can see the value of using AI to ensure that what they're how they're responding to that and what they're saying is um fulfilling their duty of care and compliant with whatever policies and standards and stuff that they have to meet. I can certainly see how that would be valuable. Um, but again, I think it's that whole thing of like you can't use AI without a person, a per like a human oversight, because it might just be like the AI spits out, like, oh, here's something you could say. Well, the teacher still needs to look at that and be like, is that actually reflective of this situation for how this student is going? And does it consider their family circumstances? You know, they might be dealing with a parent, but who knows what's going on, like two different families with like there might be two children presenting in a very similar way, but because of their family circumstances, I would imagine how the parent deals with each of those situations could be quite different. Um, and so while like say having an AI guided response might have a series of similar things that are talked about, like how that would be delivered and what specific details it goes in that you go into as a teacher would be different depending on the individual student circumstances. So yeah, I think it's really that I can see the value of using it in some of those professional situations like that, but you can't you really can't, especially when dealing with people who live in a complex world, um, you really can't use it without your own oversight.

SPEAKER_02 20:06

Um yeah, like a balanced approach, potentially and reviewing. Yeah. Well, that kind of answers my next question is where do you think humans still do things better than AI? And where do you think AI gen um genuinely helps? So it's like it's that kind of personal aspect, like we were talking about. But is there anything else that you would say humans do things better than AI? Probably a lot.

SPEAKER_03 20:35

Or creativity, maybe? I guess like in theory, like AI can be useful for sparking ideas. I've found it useful for that. Um, but I know, I guess, like from the way that it works, it pulls from things that have already been created. Like, and so it kind of my understanding just from stuff that I've read is that it kind of like replicates and reproduces things that already exist rather than kind of formulating new ideas in the same way that humans can.

SPEAKER_02 21:06

Yeah. So the authenticity is like I guess authenticity, yeah. Yep. Yeah. Um, if you could fix one big misunderstanding about AI in society, what would it be?

SPEAKER_03 21:21

Uh, I think it would be I guess the black and white perception that people maybe have of it on the whole, like maybe not on the whole, because I don't think everyone has that a black and white perception of it. But I feel like the way it's kind of talked about in general is that you're either pro AI or anti AI. And just kind of like, you know, recognize like having maybe a broader understanding of like its benefits and shortcomings and limitations and concerns, and maybe recognizing, oh, it's actually really useful for this, but it's quite probably. problematic in this way. And I guess just like yeah, being a bit less polarized about if it's good or bad, or if you're pro-AI or anti-AI. Yeah, because I feel like there's also I feel like for myself at least I'm like kind of I like to think I'm like balanced in the middle. I'm not overly pro and I'm not overly anti either, but I like I use it for some things and I absolutely hate it for some things as well. But then like I know some people are super pro AI and I know some people are super anti-AI and then I'm like oh I feel like I can't actually like disclose how I use it to some people or how I don't use some people because they're so judgment. Like I'm a bit secretive about some ways that I use it because I'm like oh this person's going to judge me so much. Or like oh this person's going to think I'm like a boomer if I don't like using it for this reason.

SPEAKER_02 22:51

Yes, that's so true. Yeah I can definitely relate to that. And I think that kind of feeds into this next question is maybe building skills to know how to use it in certain situations. But what skills do you think people especially students will need most in an AI shaped future to kind of be a bit more disconcerning but be a bit more balanced like you said and not so polarized by the completely anti and the completely pro?

SPEAKER_03 23:18

I guess like what skills especially students so I think maybe regardless of what area the students might be studying in um you know how it might be useful for their own sector and their own type of work because perhaps how it's useful for me as a researcher might be quite different for how it's useful for say like an accountant or a teacher or um a retail manager. I don't know just you know all these different jobs like understanding the benefits and shortcomings for different types of jobs in different sectors. I think as maybe more like general broad skills for students is how to be critical of the responses of AI and the outputs like I guess how to reflect on what you're seeing come out of AI and how you maybe make a decision on like oh is this a trustworthy piece of information that I would redistribute or is this something that actually requires reworking or just put it in the bin, you know?

SPEAKER_02 24:41

Yeah. Yeah. Yeah 100% because I suppose like I mean when the internet came out obviously we came up with plagiarism and certain like um programs that can detect that but you'd still have to you know you go to the internet finder um an article or a reference and you you couldn't exactly copy that you'd have to you know put it in your own words. So it's like and then there was a program to detect that do you think something like that would be worthwhile to deter like plagiarism or like I think it would be yeah I think it would be really useful um and I know it's really hard because the detection like AI is kind of improving so rapidly and the way that you detect plagiarism is quite different to how you detect AI generated text or yes things, media.

SPEAKER_03 25:37

So I don't know what it would look like, but I do think something like that would be beneficial because I can you know you hear about a lot of students like maybe becoming over reliant on AI for completing their assessments to the point where there's actually no oversight. It's just like put in a prompt for like this is my um assessment instructions and then it just spits out like a 2000 word essay with citations and stuff.

SPEAKER_02 26:05

Yeah. What do you think um I suppose students doing things like that, what do you think the impact is like for them and I guess for society if I think for them uh they're not actually learning like I mean you can you know read the assign read the essay that it puts out and be like oh yeah like now I've I've learned about the history of pandas like in China from the last 2000 years.

SPEAKER_03 26:36

I don't know like whatever the topic is but they haven't so they've learned what it has told them but I think without having gone through an information seeking process where you're like getting some facts and some ideas and whatever and then consolidating that information into your own summary I think you kind of lack that a key aspect of your own personal learning where you're not just learning about the topic but you're learning how to handle information yourself. Because I think that in itself is a really important skill to develop because like you know when you're doing a job especially a job like my job where you're always on new projects that are like different to the old ones so you have to learn new information, new skills an important thing is to like take in different information from different places so you can formulate an understanding and um you know ideas on the topic rather than just like regurgitate information that you've seen elsewhere. Yeah if I'm rambling a little bit but I think perhaps they're really they're missing that um the ability to synthesize information and consolidate ideas.

SPEAKER_02 27:52

Yeah. Yeah exactly and then you need nuance with anything right like yeah like you said you can't just kind of copy paste everything that you've learned and think that it applies to a different project in a completely different field with completely different people involved like you need to kind of be able to adapt um that information in different contexts yeah that's yeah interesting um what's one thing you hope AI never replaces um um maybe like music and movies and like those entertainment art things that we all like to consume to feel good about being alive um so yeah I think those sorts of things because of I I don't know just something gives me a bit of an ick about like artistic stuff that's not created by a person like that's created by machine. Yeah and it's like you said it kind of goes by what's already in the ether right so then you're like is this actually unique? Is this actually a credit to the person that potentially firstly you know developed this idea or whatever it is.

SPEAKER_03 29:06

Yeah and I think that's kind of an ethical thing then isn't it potentially and that kind of like that human thing as well um because like you know you think about good movies and music and like books like the stuff that you consume and like the good ones make you really feel something and you're like oh that really like I don't know touched a part of like my heart and my soul or that made me think about the world differently or helped me like I don't know see something from the perspective like outside of myself and I think I don't know I just don't my feeling is that like uh a computer code can't really like provide that for you in the same way that a person who has either like had some sort of experience that they're conveying through their creation or are sort of like conveying ideas or feelings or whatever from other stuff to you. Yeah I just I don't feel like a computer code can provide that in the same way that or even if it could like I don't know it just doesn't feel right.

SPEAKER_02 30:17

Yep yep totally um all right so we just want to do a few questions about sustainability and AI. So would you use AI less if you knew it consumes significant energy and had a negative impact on the environment I like to think that I would and I do try and be very mindful of using it um I mean I don't know what you classify as necessary or unnecessary but like I try to keep my use contained to things that I feel like I really need it for because I'm like struggling on a task at work or it's a really time critical task and I guess I don't tend to use it too much in my personal life.

SPEAKER_03 31:03

Or when I do it's like pretty minimal and mostly related to like yeah similar stuff to before like trying to understand like jargon or do like budgeting sheets for like costing breakdowns of something that sort of thing. Yeah.

SPEAKER_02 31:21

Yeah. Um would you accept slower or less power powerful AI tools if they were more environmentally sustainable I probably would because I feel like it would also still be a lot quicker than relying on like Google and then like putting together spreadsheets and stuff myself.

SPEAKER_03 31:40

Yeah.

SPEAKER_02 31:41

Mm-hmm do you um currently know um I suppose the usage or the impact environmental impact of like the AI tools that you use at the moment?

SPEAKER_03 31:53

Um a while ago I remember it being like a bottle of water for every prompt that we go to chat GPT. I don't know if they're like updated or different um environmental impact measures that are happening now but that's the last figure that I'm aware of.

SPEAKER_02 32:14

Yeah would you like supplement other maybe like using AI say it is a bottle of water and then maybe in another aspect of your life you go okay well then I'm going to try and save a bottle of water to like make up for that deficit in the sustainability or I mean it's yeah it's that like ways of offsetting yeah I guess. Yeah do you do that in other any other thing that you do with your daily life? I mean you do you talked about your car like changing to like a hybrid and EV.

SPEAKER_03 32:49

Yeah I think like for me I wouldn't say I have a uh a strict thing that I follow of like if I'm gonna do this then I'm gonna do this and if I do this and this but it's more of a general like trying to be aware of the things that I do that are less environmentally friendly and other changes that I can make that are environmentally friendly. So um you know it's I guess it's not a transactional approach but it's more of an overall consumption approach like you know types of cars that we drive um I think these days as well I'm also a lot more conscious of like um like buying clothes like I try to limit buying in like new clothes I try and do a lot more pre-lubbed clothing and just buying less clothes in general um particularly because that for me was like a big thing that I used to consume unnecessarily on. Yes me too yeah it's I think it's a common thing. Yeah and like food like I try and eat and cook vegetarian things throughout the week so not all of our meals are meat. Yeah it's more of those sort of broad shifts in my behaviors rather than a transactional approach. I'm aware that AI is like AI use is not a good behavior for that sort of thing um which is why I try and be mindful of using it unnecessarily and yeah also just make those other bigger shifts in my life.

SPEAKER_02 34:19

Yeah well done that's great. Yeah so a work in progress yeah me too definitely me too um would you pay more for a greener AI service powered by renewable energy yes but probably to a limit and I don't know what that limit is to be honest.

SPEAKER_03 34:41

Yeah.

SPEAKER_02 34:42

What would you be happy to pay monthly like do you pay for your AI at the moment?

SPEAKER_03 34:47

I don't at the moment but I think maybe for how much I use it maybe like five maybe ten dollars um just because I'm I wouldn't say I'm a heavy user of it.

SPEAKER_02 34:59

That would be like a month thing you'd be like a subtle um who do you think is most responsible for reducing AI's environmental impact users companies or governments I think all of them but I would weight it more heavily to companies and governments.

SPEAKER_03 35:19

I think governments have the scope to um that's the word I'm looking for I guess like not control but influence how both users and companies are doing it. Like maybe governments can do things that influence how users use it but also maybe put certain things in place that um limit how companies can provide the AI. You know for example the fact that kind of similar to social media in the sense it sounds as though AI is almost just like some of the available ones are designed in a way that is trying to keep the user engaged so giving it responses that they think that it thinks we'll like to encourage it to keep using it and you know to essentially be an enticing thing make make the person's life easier and make the person feel good. So I think that for example like the company has a responsibility but also they're trying to sell a product so their incentive is to get people to use it as much as possible because they make more money. But maybe then if there's government oversight and like safeguards in place on like you actually just can't I don't know how it would work.

SPEAKER_02 36:43

I don't know enough about this stuff but there has to be something maybe that um influences that like policies or like a cap. And if companies go over that cap they have to pay the tax on the usage like a carbon tax or whatever.

SPEAKER_03 37:00

Yes something like that or just like it's not allowed in this country if it's not meeting these requirements something like that. I don't know I think like you know how that whole thing happened with um like Pornhub recently and then it's like there's then it's just not available in Australia. Yeah because like of the um whatever the things that the government put in place on it like yeah the ethical yeah yeah and like kids' social media stuff as well um what's another thing I'm thinking of and it's kind of it feels kind of similar in a sense to like you know how there's been all this thing of like oh like don't use plastic straws because it's bad for the environment and then it's just like people like the population not using their plastic straws but like then there's all these like oil and gas companies doing all that stuff and like yeah now AI which is also bad for the environment like you know who is the who is really responsible there.

SPEAKER_02 37:56

Yeah that's true. And then it just seems like those things that come into place is like okay what kind of impact is this actually happening like having at the moment.

SPEAKER_03 38:04

Yeah it's nothing to tell like an individual to like not use a plastic straw or don't or not use AI excessively but if you're not actually like doing something at the big the thing that's actually creating the biggest problem. Yeah which is the the infrastructure as well right to build all these kind of um I suppose what are they called like what are they like generator systems that the data centers yeah the data centers yeah I don't know I don't know a lot about what they all are but yeah yeah I think it's like a it's a big thing but it's not just down to the individual.

SPEAKER_02 38:42

100% yeah um should educational institutions like schools universities um provide guidance or policies on sustainable AI use?

SPEAKER_03 38:54

Uh yes Anna but I do also think this kind of also is similar to the previous response I think it's really important for like schools and unis and um educational services to provide that information to like students and other consumers um so they can make individuals can make informed choices and maybe in just how they navigate their life and how they vote and all that sort of stuff. Like it's good to have that information and make decisions in line with your values. But I think that's not its own solution. I think it's like an understanding of where have I think still having a broader understanding of how that where that responsibility should lie and yeah.

SPEAKER_02 39:44

Yeah definitely I suppose that that lends into the next question which is what role should educational institutions play in modeling or teaching sustainable AI use to students but specifically to students I suppose like policies and things in at a university level or whatever. What what role do you think that they we should play? I suppose it's more about as academics or as you know clinicians or whatever how do we kind of model that behavior if we're using it quite a lot and then we're telling students not to use it I suppose that's a bit of a discourse of what you know practicing what we preach.

SPEAKER_03 40:25

Yeah that is true and I that's actually something I reflect on a bit myself is like oh you know I feel frustration frustration with the thought of like students using AI to complete an assignment but I'm still using it to do certain tasks for my work. I like to think that I'm using it responsibly. I guess is my concern is more like if students are maybe not using it responsibly where like you know like I said before this random example which I don't know if I've encountered it or not like where students just put in the a prompt for the assignment instructions and then just submit it whatever it's spat out like as is um yeah I think I think it probably you know given how new AI is I don't think anyone has really nailed a great way of teaching and modeling like good sustainable AI use yet. I think as a society we're probably in the figuring it out stage. But yeah I think I think it is important for educational institutions to you know you can't just you can't just tell people not to use it because they're going to use it and there's no way that you can actually stop them. And right now there's not actually any way that you can detect that accurately and like consistently so I think it's maybe looking at how you might bring AI into the curriculum kind of like once upon once upon a time when like Google searching was first brought into the curriculum and it's like oh this is how you search. I remember in school like yes in IT or whatever it was it's like oh this is Google like this is how you search things in Google and like when you put in your search term it shows up all these pages but then when you click on the pages they're all different and like you have to figure out a way of like knowing what is good and what is not good. I think it's maybe in a sense something similar to that that we might need to do. Yeah and yeah I think an example might be I mean I I don't know I don't teach so I've I'm not coming at this from an informed angle but like incorporating AI into assessments but maybe also getting students to explain how they used AI in their assessment piece and what they did with the prompts. So it's kind of like stepping them through that critical thinking process. Because while they're students like that's when they're kind of forming their knowledge and mental skill sets to do whatever job in that like that they're going to do eventually. So yeah yeah that's a really good way to approach it.

SPEAKER_02 43:12

Yeah because like I know you might have some research students that you supervise do you how do you approach AI with them or how have they have you s potentially had any work given to you um that is a I guess more of a like higher level research as like an honours masters or PhD um you know that you have to say oh this is clearly AI and I would prefer you not to use it. However I know as academics obviously for writing wise it's a gray area. It's a very gray area when writing manuscripts and things I know some some do it some really use AI to formulate you know their manuscript and some don't or it's a bit of a bit of both. Um, so how do you approach it with like a research student where you know authenticity of their work is part of the criteria they get marked to get their degree?

SPEAKER_03 44:10

Hmm. I think um I haven't done a lot of like uh research degree supervision in the past couple of years. So I don't really have a specific example that I can draw from. But how I like to think I would deal with that is I would try not to have a hard, fast rule on like you're not allowed to use AI because one, it's not like I'm not using AI, like I am, and two, if I tell them that, like me telling them that isn't actually going to stop them from doing it. Um I I feel like I would try and establish a open environment where they can disclose, like, oh, you know, I used AI for this. And then like my way I would like to teach them is like, okay, you know, you can use AI. Let's look at what sort of output it's provided and what we actually need to do with it to ensure that it is still your work, like you're actually evaluating the quality of what it's providing and you're like ensuring that it's addressing what you need it to address. And I think just like, yeah, just making sure that it's still your ideas, it's still your work. Yeah.

SPEAKER_02 45:35

Yeah. Just creating that. Yeah, because that's the thing. I think of like me as a supervisor, even as a future supervisor, how would I approach um maybe um some areas where I think oh using AI for this would be really great. Like you you you can use AI for this because this would actually save you a lot of time. However, for this, no, no, no, AI needs to be it's yeah, and I feel like that for us is quite clear about you know, if I I feel like it's a slippery slope potentially, especially with new students coming in. Yeah, so maybe like we were talking about, is that the uni, especially in research students, needs to have some clear guidance and structure around that. Um because I've got some people I work with that are like absolutely no AI is used for anything. Like at the front, it's like nope, not except we're not using it for anything because so then it doesn't get grey later.

SPEAKER_03 46:28

Yeah. So yeah, it's interesting. Yeah, it's interesting, and it's really tricky. And I think like um, like you said, using it for some things because it will save so much time. And I think like maybe in the context of a like a research degree, it's understanding what um aspects of your work are more important to do without or with minimal AI input. So you can actually develop that skill and that thought process and that ability to like generate ideas and synthesize information and communicate well, versus like, oh, systematic review screening. Like you do, you do maybe like half an hour of that, and you're like, okay, I've I've got an understanding of how this goes. But like having to then spend like up to 20, 30 hours doing screening is like it's not great use of time. It's definitely like the sort of task that A and I could help make faster.

SPEAKER_02 47:24

Yeah, so true. And also like thinking of it as it's a training course. So we need to give them the skills at the outset, and then you know, now that we've developed that, it's just I feel like it's so challenging with students coming in at the moment, you know, reset students. But yeah, it's definitely something to keep chatting about. But um, thank you so much for chatting with me today. You've been amazing giving so many valuable ideas about this area, and yeah, coming from such a unique lens as well, I think it's going to be really useful um for people to listen to. So thank you so much.

SPEAKER_03 47:57

Thanks for having me.