SPEAKER_02 0:00

To start us off, would you be happy to share your current role and area you work in or study in?

SPEAKER_01 0:07

Yep, so currently I'm studying and working in research. So with my study I'm completing a master's of research that's looking at trust in AI systems in relation to health. And then in my work aspect I work as a research associate on a project where we're developing AI coaching software and navigating what's required as far as user perspectives in developing AI tools.

SPEAKER_02 0:43

Yep, awesome. So this is so bang on with our AI topic. Yep.

SPEAKER_01 0:47

So that's awesome.

SPEAKER_02 0:48

And you also have a background in nutrition too.

SPEAKER_01 0:51

I do. I do my bachelor's is in nutrition and exercise science. So that's where the health component comes in.

SPEAKER_02 0:58

Yeah, incredible, amazing. So when you hear the word AI, what's the first image or feeling that comes to mind for you? Anything? Um productivity, probably. Yeah, that's a good one, 100%. Yeah. What's one everyday task you happily hand over to AI and one you never give up?

SPEAKER_01 1:19

Um something I'd happily give to AI is just tightening up wording structures in any most texts that I produce. Yeah. And probably brainstorming is another one. Um AI to get ideas, um, it can be pretty useful. Um and then something I wouldn't give up to AI, um, obviously, I think putting in um any kind of really detailed information, sensitive information, um, so when you're um trying to put together something that's a little bit more personal or has other people involved in it, then yeah. Trust in AI is definitely an evolving um issue, I think.

SPEAKER_02 2:05

Yeah, definitely. And would you say that would come under like writing personal emails or personal messages to people, like things like that?

SPEAKER_01 2:12

Depends what the context is. I think if you give if AI is um writing an email for you and you've given it some context, but there's actually no personal details in there, or there's not phone numbers or like sensitive. Specifics, right? Um if it can draft you up the email without specific details and then you use it how you wish. Um I think that's probably okay, but yeah, just that yeah, that involvement of sensitive information is probably probably the hard line.

SPEAKER_02 2:42

Right, like confidentiality. Absolutely. Yeah, yeah, yeah.

SPEAKER_01 2:45

Yeah.

SPEAKER_02 2:46

Yep. Um what's the sustainability habit you've picked up recently? Sorry, this isn't focused on AI, this is just generally about sustainability. But yeah, so what's a sustainability habit you've picked up recently, big or small, can be anything that you maybe swapped over to be more like conscious of the environment or yeah.

SPEAKER_01 3:07

Um probably the thing that comes to mind first is um using our scraps from the kitchen and chucking them into the garden. Yeah, like composting. Yeah, so we usually well we used to just um chuck it straight into the um green bin, which I mean is composting as well, I guess. Yes. Um, but now just chucking it straight into our own garden to benefit our own garden.

SPEAKER_02 3:30

Oh great. Is that like you grow veggies and fruit and stuff? Ah, amazing. Do you find like animals come and like eat it or you have a problem with pest stuff or no?

SPEAKER_01 3:40

Uh not really. We have um our garden has a lot of native plants in it, so we do get a lot of birds anyways, but I don't think they're destroying our vegetable garden too much.

SPEAKER_02 3:50

So yeah, yeah, it's fine. Or coming in and eating it, maybe things that they yeah, yeah. I mean, I guess it's just fruit scraps around.

SPEAKER_01 3:57

That's right.

SPEAKER_02 3:58

If they're gonna pick at it, it doesn't matter. Yeah, yeah. Oh, good on you, great. Um, so back to AI, um kind of topics. Um, when you think about AI in your field, what's the first thing that excites you or worries you, or either? Either way.

SPEAKER_01 4:16

So I think um AI has got great potential to um be really helpful in a lot of a lot of different fields. Um in health specifically, um, it can probably help with um workflow practices, um taking notes, um, clinician notes and things like that. I think it's got a lot of potential in areas like that. Um again, accuracy and trust comes into that because you're working with sensitive information in health.

SPEAKER_02 4:50

Um yeah, yeah, no, that's a good point. And yeah, I suppose that would just be a line that you would have to like consider when you are doing like clinical notes and things that it it needs to be kind of de-identified in that way and checked through. Absolutely, yeah, absolutely. Yeah. Um so that's I guess probably like one of the key kind of things that worries you and that's like confidentiality again.

SPEAKER_01 5:15

And I think the confidential confidentiality um and the accuracy um AI has a tendency, as most people know, to hallucinate, to make things up. Um you can um convince it of something that isn't true and it will then tell you that back to you. Yeah. Um so um you can give it structured or numerical information and it with really clear guidelines on what the output is, and it can still get that wrong. Um you could you can put the same um prompt into it twice and it will give you a different response each time. So um yeah, there's a lot of um inaccuracies in AI, in using AI, and I think the user probably needs to um be really aware of that.

SPEAKER_02 6:13

Yeah. Um have you had any like examples that you found that it was completely like wrong in something that you put in with a prompt and you know could have been really problematic potentially if you didn't pick it up or um there's probably been a lot, but I can't think of one off the top of my head. Yeah. Yeah, feel you. Because even sometimes I'm like just some easy task, like, oh count how many rows this is so I don't have to and then it will just make a number, and then I'll double check it and it's completely wrong. Yeah. And then you tell it that was wrong, and it says, Oh, sorry, and does it again. Yeah. And it's still wrong. Yeah. So yeah, it's quite interesting, isn't it? Um, how do you explain AI to people who aren't in your area, like students, colleagues, family, and what do they usually get wrong?

SPEAKER_01 7:01

Um, so I have a really great kind of example of this. Um, my dad is actually staying with me at the moment and I was trying to explain to him that I work in the AI space and he'd not ever used it before. Um, and I said to him, it's kind of like Google, where you can like ask it a question or um look something up, and but it will give you a much more detailed response. Um so yeah, I think that's kind of like how I would explain it to someone, it's just like looking up information, but the information may not be correct, but then I think Google is the same. Yeah, Google may not be correct and it may not direct you direct you to the web pages you're actually trying to look for. Um so while AI can be incorrect, the internet can also be highly incorrect. Yeah.

SPEAKER_02 8:00

So that's true. Yeah. What was his response?

SPEAKER_01 8:03

Like, was he like, oh okay, or like, oh maybe I'll use that, or oh that doesn't interest me, or um he was actually really interested and he uses it um uh using the voice model, so he talks um to his phone, yeah, um chat GPT on his phone um and asks a bunch of different questions, I think, just to like see what it will answer, like see what he'll get out of it. Yeah. Um he thinks it's just fun and interesting and yeah.

SPEAKER_02 8:29

Oh cool, there you go. Do you know what he's what he's using it for?

SPEAKER_01 8:33

Like, or just he was asking it um about um because he has diabetes, so asking it about how does it how he can control his diabetes and um what foods he should be eating. Oh interesting, right.

SPEAKER_02 8:47

And when he was doing things like that, did you feel that he was like, I'm just gonna take on that information, or was it, you know, did the did Chat GBT say advise this with your GP or a special?

SPEAKER_01 9:00

I'm not really sure um what kind of level the response went to in regards to like advising him to ask at like his GP or whatever. Yeah. Um but the responses that it gave I felt were fairly accurate. Yeah. Um, and I did say to him, like, you do need to double check what it tells you kind of thing, just because I do have that background in AI and have a good understanding of it's not usually 100% correct.

SPEAKER_02 9:29

That's interesting. Um what changes do you think AI is going to bring to how we learn, teach, or make decisions?

SPEAKER_01 9:37

Oh my goodness. Um I think in terms of learning, AI has a really great capacity for breaking things down to a level that meets the user. I have for an example, um I have a daughter, she's in year seven, um, and she came to me with a question about long division, I think it was. And I said to her, hmm, I can't remember how to do long division, we're gonna have to Google that.

SPEAKER_03 10:09

Yeah.

SPEAKER_01 10:10

And we actually ended up taking a photo of her question on a worksheet and uploading it into ChatGPT, and we asked ChatGPT to step us through how to answer this question. Um, and we just left um this microphone on on ChatGPT, and as she was working through the problem, she was if anytime she said, Oh, I don't really understand that bit, ChatGPT would jump in and break it down for her into even finer steps, and we got through the whole long division question. Um, and I said to her at the end, just write at the bottom of your worksheet, that we use ChatGPT to help us answer this question. Um, so that we're being transparent about how we got there. Um, I would have had to Google it anyway, so yeah. Oh, but I think the capacity for meeting someone where they're at is huge and having the time to break it down for that individual is really important. Thinking in terms of a um like in a classroom environment in education, um, primary school, high school. Um a teacher has only a certain amount of hours in the day and has so many children that they need to attend to. Um, so yeah, I think in that capacity, AI has a huge potential to be able to help students at that individual level.

SPEAKER_02 11:35

That's great. And I suppose it's the same for parents and families, right? Like, you know, you've you've got other children, you've got other things, so like spending time doing helping with homework and stuff might be out of your like for me, out of my death completely with absolutely also, yeah, so it's a great tool to use potentially like with someone else to work through.

SPEAKER_01 11:57

Yeah, yeah, yeah. And even like um it can help you learn languages now, like and have an actual conversation with an AI um in a live learning environment to learn another language. That's an another huge capability that's true, could be really useful in the future.

SPEAKER_02 12:18

Yeah, that's a really good point. Yeah, I like that idea because I was thinking about learning a language. You just have to trust that the AI is translating correctly. That's true, that is true. Um, where do you think humans still do things better than AI? And where do you think AI gen genuinely helps? Which I think you've touched on, but what do you think humans still do better?

SPEAKER_01 12:44

Um humans definitely have um the emotional and critical analysis skills that really um are better than what AI currently has. Um with my work, we are working on um training coaching um AI bots. Um so we try and train them to be um have empathy and have an emotional response. Um, and it's really hard to get it to do that. Um you have to be really specific in your prompting of an AI to get an emotional response out of it. Um, so yeah, humans definitely have the upper hand there. Um and it in terms of being critical, an AI response is probably only as good as the prompt you give it. Um, so it can only base its response on what you've put in. So if you don't give it enough context, it can't give you a critical response back unless you've given it everything it needs to make that critical response.

SPEAKER_02 14:09

Yeah, 100%. That's so like it's challenging with prompts because do you think that they're I mean at the moment, obviously you're working on some kind of prompts that you can put in that will give you a similar response every time, depending on who answers it and their context, that it's gonna ask the right questions to get the context, right? Is there currently like any support to develop prompts at the moment, or like depending on what you're wanting to get from it, is there something like a I don't know, gold standard kind of I don't know, template or framework or anything like that?

SPEAKER_01 14:47

I don't think there's like a gold standard framework. Um I know that um software such as ChatGPT and Claude do put out um guidelines around um how to prompt. Um, there's definitely blogs um on the internet around um creating a good prompt. Um there's probably been research done on this. Um not that I've looked in too much into what research there is on actually prompting the AI, um, but it can very much be trial and error. Um yeah. And the model of whatever model you're using comes into play as well. So as we know, ChatGPT has a few different models within ChatGPT, and then there's Claude and Gemini and all these other AI agents, and within those they have multiple models, they all perform differently. So depends on you could put the same prompt into each of those, and you'll get a completely different response from the different models.

SPEAKER_02 16:00

Yeah, it needs to be specific.

SPEAKER_01 16:02

So yeah, so what if you create a prompt um specific for one model and then someone tries to use it in a different model, it's not going to perform the same. So just being aware of that, I think, is also really important.

SPEAKER_02 16:16

Yeah, definitely. Yeah, good point. Um, if you could fix one big misunderstanding about AI in society, what would it be? Because I think um AI can be quite polarizing, right? And how people view it. Some people absolutely hate it and completely like, nah, I would never use it, never like you know, if my friends and family use it, you know, I'm not that's you know something I don't align with, blah blah and then you got that opposite end where you know people really think it's great, and you know, so yeah, what do you think are some big misunderstandings about that polarisation between it's like should be used for everything and it's amazing, between like don't use it at all and it's terrible.

SPEAKER_01 16:58

Um I think that everyone's probably entitled to their opinion on it. Yeah. Um I think because AI is fairly new and is still evolving and evolving really quickly and will change a lot in probably the next year and then the next five years, it'll be completely different again. Um I think obviously I'm a big adopter of AI and I'm studying it and working with it every day. Um, but then I do have know quite a lot of people that don't use it and don't want to use it. Um probably most of those people are either um older people um or they've just not tried it. Um in my experience, people um young 20, 30, 40 year olds, um, probably their hesitancy is more likely because they haven't tried it and they don't know how to get started. Um the other probably the other big component of people not using it is um the environmental impact.

SPEAKER_02 18:14

Yeah.

SPEAKER_01 18:14

Yeah, yeah, definitely.

SPEAKER_02 18:16

And um do you think in terms of why people don't use it and potentially because they don't know how to, do you think it's because also they have this stigma about AI and it could be environmental, could be I've I guess I've heard a lot on the media in the media and things like that, that AI is replacing our critical thinking skills, and that you know that means that we're not thinking for ourselves, and you know, learning is impacted. I mean, from your I don't know if you've seen any research in that space, like is there actually I mean, maybe we don't even have much evidence at the moment to suggest that it's impacting?

SPEAKER_01 18:53

I think there's probably not much evidence, and I think that it's just gonna be a different type of learning. Yeah, so um a lot of um social media and news reports will say that um university students are writing their assignments using AI, and I think it depends on how they are using the AI to write the assignments. It that's the important question. If they are just putting in instructions and taking whatever the AI response is and putting that into an assignment and calling it a day, that is completely different to evaluating what the output of the AI was, checking the facts, checking it against evidence and research, and then critically thinking about how they can use the information the AI has provided to contribute to their assignment. I think if there's two different ways of looking at it, and I think AI is probably not going anywhere, and we need to adapt to a new thinking about how can we incorporate it into our learning rather than just flat out rejecting it because I don't think that's going to happen. Um and it's more about teaching people how to use it effectively and critically in their study, in their work so that it's not being abused.

SPEAKER_02 20:30

Yeah. And do you think responsibility it is to teach the next generation or to teach us as you know um you know, academics or anyone how to use AI, I suppose, ethically and responsibly. Like is it parents, is it teachers, is it institutions integrating it into policies? Whose kind of responsibility is it to, is it on the individual to I think it's everyone. Yeah.

SPEAKER_01 21:00

I really do like children will learn respect ethical values from their families, from their teachers, from laws and policies that are in place in their community, that's they will learn their ethical values from all of those places. Um so anyone who um can provide advice in an AI space, whether it's parents who have knowledge on AI, teachers or anyone else that can contribute to that AI learning, I think it's a combined effort. I don't think the responsibility just can be put on one person or one organisation to be the sole instructor on how AI is taught.

SPEAKER_02 21:58

And do you think at the moment there's enough support to know, I suppose, to teach or model that kind of behaviour in it in any of those people as an individual, as a parent, as a like, do we have enough support around us in our institution or in our sector or to actually know how to how to do that, how to support? I mean, like you said, your your your your dad was like, I have no idea what AI is and some people just don't use it. And then it's like so if we don't we don't know what we don't know it's tricky yeah.

SPEAKER_01 22:30

Yeah I think um it's probably too early to say right because there's not any one space where you can learn how to use AI. Yeah which is probably like anything that you're learning how to do these days most people if they're trying to learn something new they'll go to the internet and they can draw information from anywhere on the internet. AI is no different so I think when you're learning when people are learning how to use AI and how they should be using AI ethically it comes from taking information on how to use AI taking information on what your ethical values are putting those two together with guidance from your peers and from your support networks. Yes yeah I think it's there's no one true source for learning about using AI.

SPEAKER_02 23:36

I think yeah it's too and there's and because it's still evolving and we'll be I think for a long time it's really hard to just put an individual on that like true and then it's like in some context you don't have a choice to use AI and it's integrated into Google searching it's integrated into my emails you know suggesting do you want me to summarize this email for you? You know it's seems like it's pushed in pushed in our face so then yeah it's just it's an interesting space in that regard how it's not something that's like an the internet you can choose to go on the internet and search for something but I can't choose then to use AI if it's integrated.

SPEAKER_01 24:19

Absolutely what do you think about that as it being kind of really already integrated then and we're kind of on the back foot potentially we probably probably are a little bit on the back foot especially um in the Google space in the Google search engine when you get that AI first response. A lot of people still don't realize that's an AI response.

SPEAKER_02 24:38

Yeah oh yeah some of my friends who are like I'd never use AI I'm like have you used Google lately? Oh yeah I have yeah and you know how it comes up with yeah it does yeah yeah yeah and they're like oh I wouldn't you know I'm not going to use Google anymore and I just think well there's actually so many aspects of AI that you don't realise. Is there anything that you didn't realise was AI until later like that it uses some sort of artificial intelligence like I mean Google Maps I didn't really realise it uses artificial intelligence to estimate how long you know it takes to get to point A to point B based on other aspects but probably haven't thought about it too much.

SPEAKER_01 25:16

Yeah just I'm at the point these days where I just expect that everything's most thing most applications that I use have it in some way.

SPEAKER_02 25:26

Yeah so it's kind of yeah it's very interesting.

SPEAKER_01 25:30

Okay what's one thing you hope AI never replaces ooh I don't know I think I think this is probably a little bit controversial but companionship because I know that some people already who use AI for companions um if they are experiencing loneliness um or need mental health support but in the wider context I think that companionship should be human.

SPEAKER_02 26:05

Yeah yeah no I completely yeah I can yeah I've thought about that too and I suppose then I think about other contexts like re you know regional context or if you're quite remote um and potentially people chat online to people too yeah um but I think it's a different thing right because you're putting all this pressure on something that doesn't have that emotional capacity like you mentioned before. Yeah it doesn't have those skills so is it just going to agree with you on everything we don't know right exactly that's that's actually really concerning right like very concerning could be really dangerous absolutely to have those conversations with something who's going to agree with you if you wanted to do something harmful to yourself or to other people. Yeah that's scary.

SPEAKER_01 26:55

That's actually something that we're very conscious of um in training our aim our AI models that I work with is putting in security sorry safety and boundary prompts where if the user mentions any kind of self-harm or medical emergency the AI is triggered to move out of what it was doing and to um assess with the user if it's an this is an emergency situation.

SPEAKER_02 27:29

Interesting that's great. Yeah wow so did you figure that out just from having those slight impromptu like just testing it or is that something that it advises through doing anything with health in AI like uh no so we developed that on our own we didn't there's not really any guidelines on how to build AI and health um that's why we're doing research yes but yeah uh we had to think about what conversations are people going to be having with the AI in a health space um and mental health is a big component um so how do we tackle that and how do we make sure the AI can handle those conversations?

SPEAKER_01 28:21

Yeah 100% because the aim of the app is to provide general advice is that right the one that I'm the app that I'm working on at the moment is um general health advice around nutrition and and physical activity um to improve people people's lifestyles um but obviously people we don't know what questions users are going to ask and they could relate to things like sleep or alcohol abuse or mental health issues um drug abuse sexual abuse so um we have to have contingencies in for those questions yeah yeah amazing um so now this is these questions are going to be more focused on sustainability and AI um would you use AI less if you knew it consumed significant energy and had a negative impact on the environment so my thoughts on this uh as an individual my impact is probably not very great I think in a corporate or business environment that's where the real impact happens with AI and all these businesses and corporations that are implementing at a really high level that's where the impact is really felt as far as sustainability environmental impact and on the flip side of that I don't think AI is going anywhere and I think if you don't use it you're probably going to fall behind in the tech space at the moment so it's it makes it really hard to make a decision on that.

SPEAKER_02 30:00

Yeah and also like we said it's in it's integrated already so that's right that decision's kind of taken away from you absolutely especially yeah even uh like where we work at the moment like it's already yes yeah um would you accept slower or less pack powerful AI tools if they were more environmentally sustainable um I'd probably accept a slower tool if it was more environmentally friendly.

SPEAKER_01 30:30

I think a less powerful tool um would compromise accuracy on AI responses um and the whole point of what AI can do so probably I wouldn't compromise on performance because otherwise you are getting a less a lesser product. Yeah like a less accurate product right absolutely and when you're talking about education or health that's really important.

SPEAKER_02 31:03

Yep 100% um would you pay more for a greener AI service powered by renewable energy if it if you know you knew about that.

SPEAKER_01 31:14

Personally I would um but not everyone has the capacity to do that. So yeah. Who do you think is most responsible for reducing AI's environmental impact users, companies or governments um I think the developers so not necessarily any of the users I think the most responsible person is probably the developer of the AI then probably seconded by governments and corporations that use it. I think the individual user is the lowest ranking on that scale.

SPEAKER_02 31:52

Yep. And in terms of what do you think is needed to I go I guess entice companies to be more sustainable or like you know the developers to be more sustainable if it's going to cost them more money to maybe you know solar panels or to have water stations that are you know yeah different kind of ways of I think that's a really hard question because the world is powered by money and these companies um are often built around their profit line right yeah so um and at an individual level we probably don't have much say in how they run their companies other than to not invest in them as far as the stock market's concerned. Yeah um but yeah what would you like to see I suppose in place to make companies more accountable what do you think would be a solution or maybe a win-win for everyone I don't know if that exists yeah I don't know either um I probably don't know enough about how these systems are developed um and are run to comment on that.

SPEAKER_01 33:23

Yeah um yeah but I guess you could compare it to other types of um industries that run that use a lot of resources um that continue to use a lot of resources even though their impact on the environment um can be detrimental?

SPEAKER_02 33:47

Yeah definitely um should educational institutions schools universities provide guidance or policies on sustainable AI use?

SPEAKER_01 33:58

Absolutely yeah yeah do you think how do you think that that would be enforced or managed do you think that's a really hard one again to monitor um because I think that um maybe if schools are providing an AI service limiting the amount of turns or tokens um that the AI uses per user um could be a good way of limiting the use of the product.

SPEAKER_03 34:38

Yeah.

SPEAKER_02 34:40

Yeah that's a good one. And I suppose I don't know about like your daughter you're saying who was using the AI to help with the math problem. Do does the school from your knowledge have their own AI platform or like provider that's unique to the school or because I've heard I've heard about that but I wasn't sure it's speaking from maybe your experience that they have their own sort like I think Carol was mentioning it about the having their own platform that they use.

SPEAKER_01 35:15

I think they do but I don't know much about it. Yeah yeah do you think that would be something like then maybe I don't know institutions would have their own specific one that then can only provide or you know they would have a subscription or whatever it is and then they could make those decisions a bit yeah absolutely I think if um schools are going to implement AI into their work practices especially when it comes to involving students then those AIs need to be trained in a certain way they would need to be housed in the school's own environment the topic of data sharing and where the conversations with the AI, where that information goes um what it's capturing that all comes into play and that's what we look at a lot here at the university when we're developing AI bots is that data privacy and where and how the AI is trained. Yeah.

SPEAKER_02 36:27

And lastly what role should educational institutions play in modelling or teaching sustainable AI use to students so maybe as a student yourself you know in the university context or you know having children that attend school yeah what role do you think educational institutes should play in kind of modelling sustainable AI?

SPEAKER_01 36:53

I think it's probably going to be about being transparent on AI use. So I think in universities we know that universities are using AI they need to be transparent with the students that are attending that they are using AI because at the moment students are expected or are being asked not to use AI in their work at all yet um the P the universities that are creating the courses are using AI.

SPEAKER_02 37:24

So you've been told that to not use AI at all in your yes so when in my studies not even for certain things or like crediting AI.

SPEAKER_01 37:34

So so when you it depends on uh it's uh each assignment's different but the majority of assignments it will say um AI cannot be used in this assignment. Interesting um how do they monitor that how do they know I think in the policy somewhere it has something to do with it's based on the um lecturers or the um tutor's um judgment. Right um because I think we've learned that AI tools that are supposed supposed to detect AI um are not 100% so then that cannot be used as a guide for students work I think. Yeah.

SPEAKER_02 38:26

Do you think it's important that those tools that detect oh sorry I don't detect AI are like valid? Like do you think do you see value in actually developing software that can detect AI use?

SPEAKER_01 38:39

I think in the future it's probably not going to be necessary because I think that the use of AI will become more widely accepted. It'll be more about teaching people how to use AI rather than it like it will be expected that people will be using it.

SPEAKER_02 38:57

Yeah yeah and in terms of yeah sustainability around that what's your thought about I suppose is there any anything mentioned in those policies about you know not only using AI in terms of um I guess IP or you know but also there's no is there any mention of sustainability not to my not to my awareness interest yeah I think um mo in a university student perspective um the AI the advice for not using AI is purely based on generating your own work.

SPEAKER_01 39:40

Yeah um I cannot I can't recall um any statements on sustainability within those policies. Interesting not that I've read them in high detail.

SPEAKER_02 39:51

Yeah interesting I just think of you know um I don't know if you used to see people's emails that would say think twice before printing this email. Yes you know like for like environmentally think twice before printing you know I just like a really good point. Yeah I just find that like interesting in that context of like just printing an email is that you get a bit of a shame maybe chat GPT needs a disclaimer at the bottom think twice about submitting this prompt. Yeah do you think and then it could say you know you might be using two bottles of water to all of a sudden it's a prompt. Do you think that would actually change people's behaviour?

SPEAKER_01 40:25

I don't know it could it's interesting isn't it yeah a lot of people probably aren't aware of what um is used like the water that is used um in AI.

SPEAKER_02 40:39

Yeah because I don't even I mean in terms of actual evidence about what is the usage like do we actually have good data on that I don't know. So that may be a s a move as well potentially oh amazing thank you so much for sitting down with me and chatting. Is there anything else that you wanted to add that I didn't touch on or that you wanted to talk about? No that was great