 Morning everyone, so this is going to be fairly fast-paced session because we had a really overwhelming response. Thank you so much to our panelists for joining us We've got three prompt questions. We're going to go through we're going to give two to three minutes for each of the panels to answer Those questions and then we'll have audience questions towards the end. So that's the general format for this Would the panelists please introduce themselves? I think that would be the easiest way to do things my friends on This is a large from very efficient. This is a Business in making online proctoring solutions and I grew up in India where The importance of exams which is very very high because of the competitive nature came to us brought my grad study and did PhD in computer science in general building large-scale systems and I was asked to Teach remote in South Korea where I was Proctoring a student when the student was taking a distance course from Georgia Tech in Atlanta So that's where I felt like there's a need for a better way to do online proctoring. This was like 13 years back and Later on I basically Started using AI to how to improve the exam integrity in for online learning. Yeah Thank you Good morning. My name is Megan Menser and I am here. My business is harnessing your potential. I'm a functional health practitioner I got into functional medicine Because I struggled with auto immunity myself They told me that you would always have auto immunity and after 10 years of you know Learning myself. I'm auto immune free and I've just been on a journey to really just help other people Via functional modalities if you don't know what that is functional medicine is all about getting to the root Not just symptom management and how kind of AI works into that The method that I use is a four-step method and it is called the hope method So really what I'm on a mission to is to help people harness their potential to Revolutionize healthcare and spread hope to the world and a lot of the projects that I've been working on I also do health consulting with some hospitals and we've been utilizing AI and kind of a healthcare setting so just talking about even you know yesterday we talked about if we would have personal assistance and Nor health will kind of has some elements of that as well too. So very excited to see the future of that Hi, some of you know me I'm Elizabeth Dalton currently with Intellibord. I've been in educational technology for about 30 years and At Intellibord what I'm working on as a learning and development analyst is gathering data from all the different kinds of technology that learners connect with and Bringing it to one place and focusing and manipulating that data or interpreting that data to make predictions about learner success There are a lot of different definitions of learner success. I'll be talking about those in my session at 1245 and how I'm spending a lot of time right now experimenting reading thinking about how new emerging AI tools can help us To learn to understand what learners are doing to understand how learners are affected by their learning environments by the people that they are learning with and To be able to gather that data and I mean we're talking big data right lots and lots of data And this is where I think AI can really help us to interpret that mass of data and use it for human purposes to help us reach our learning goals and To help the learning process be more available to everyone So that's what I'm hoping to talk about today Hello, I am Tim Hunt. I am a senior software developer at the Open University I suppose if I got to pick my job title, I would say educational software developer So professionally I'm a computer programmer But I would say as an amateur. I'm a very interested spectator on the kind of Scholarship and science of how education works and how a technology supports that So hopefully I have something to contribute today Hi everybody, good morning. I'm Heike Vilennios from the University of Helsinki. I'm an e-learning specialist there I'm also a postdoctoral researcher at the same place and I'm just starting a new research project on how AI tools are changing the workplace, especially the work of software development So I'll be sort of trying to wear two different hats to this discussion today So everyone just like to say thank you to our panelists for joining us today, and it's absolutely fabulous to see everyone here So the first question this has been a big one in the media and with a lot of the a lot of the dialogues around AI So really data representing the world as it stands with its existing biases is used for training AI and Perpetuate some of these biases and the impacts and has an impact on students The question that we want to discuss now is do you think educational institutions institutions are equipped to be prepared to address that? Potential AI bias now. I've chosen images for each of these questions. This is an image of a Senegal Barbie, which was in a box click article a little while ago May not be immediately obvious what the problem with this image is she's carrying a k-47 that that articles pulled down very very quickly So we want to be able to understand and I'd like the panel station discuss What do they think that institutions can actually deal with these issues? Is it their responsibility or do we think it should be pushed back to the vendors on that? So Elizabeth would you like to start since you're a data specialist? Well Are our educational equipped institutions equipped or prepared to attend Address potential AI bias. They're not addressing potential bias without AI so No However, it is not all grim We we can start to use AI tools to look at bias and Again, this is part of knowing that we have a lot of data If we expand so that we're getting data from more different sources More open sources, we're you know at an open software conference One of the problems that we have with a lot of AI tools right now is that they are trained on the you know weird data set white Western higher ed wealthy and What we need to do is be including more data sources more voices and Then we will be in a better position to be able to look at bias but there has been some very interesting research about the bias and We're starting to see for example in the area of learning analytics one of the more interesting papers that I saw a couple of years ago was If you've ever done anything with predictive learning learning analytics There's this graph that you end up seeing the receiver operator curve So you've got some grid and you've got a diagonal line And you've got an arc and you want that arc to be pulling away showing that it's making more and more effective accurate predictions Slice that data by your demographics and say does it equally accurately predict for all your different Subpopulations so people talk to me about will it be ethical to include demographic data in predictive learning analytics? Won't that introduce problems? We have to include that Because we have to be able to analyze our AI systems by that demographic data or we will not detect the bias So this is where you know We'd all like to be colorblind and wealthblind and all of that but that's an illusion to a large extent We can try to be that in our day-to-day lives, but when we look at the data. We can't pretend We have to actually look at the hard messy details and say if we're going to try to get bias out of our systems Or we're going to at least acknowledge where our systems are biased and they always will be biased because they are created by human beings And we are biased by our own experiences It's the way our brains work But we can try to overcome that and we can try to help our AI systems overcome that by taking a good hard look and Testing our systems against those biases. Thank you. That's brilliant. Who would like to speak next? Yeah, I think when I contemplated this question just emotional intelligence always comes up for me So playing on that is obviously human beings. We were created where the sum of our experiences And I think as educators and even if the the data is biased, correct We should utilize that to say hey, this is a beautiful opportunity Where are opportunities and how can we grow and develop and how can we make it more accessible? So how can we use this tool? How can we develop those critical soft skills? How can we develop more emotional intelligence and understand people from areas that maybe we don't have that experience but With AI and getting a more broader view. I think that we can actually address those biases More by just understanding our own human nature Like a background kind of an anthro too and I think studying ourselves is always the most difficult so it's beautiful to be in a place where there's people from so many different perspectives and Having more open-ended conversations to seek to understand to assume that positive intent and utilize this technology and just understanding that like This is the direction. We're going we're growing and we're developing and It's a great tool. We just need to be responsible. So I think in the education system we should embrace it and Just use how we can most responsibly Utilize it to grow Yeah, thank you Megan and thank you Elizabeth. So one thing I disagree is that We are focusing on bias in the way where it creates more fear and the advantage of AI Will be limited if adoption is not there So from that perspective like to bring the difference between ethics and bias. So sometimes bias is needed So one example, I like to share is let's say you are building a predictive model for Finding out at-risk students in a classroom to be able to help them early in the course, right? So earlier in the course you have limited data real data about the students and and how they are participating in the classroom So if you don't have enough data and is still you want to predict Then the data model you will choose Will need to have some bias to be able to Accurately figure out early enough so that you can intervene early enough So from that perspective sometimes bias is actually good What is more important than bias is ethics like? ethics is basically the what is the principles what is the intention behind building an AI system and And then based on those ethics we have to implement Accountability that whether the system which is built is following those principles or purpose That was put in in the beginning and so I like to share that When we are building a system before we worry about bias let's really focus on purpose and ethics Then set up controls in place to make sure that we are really intentionally moving in the right direction and By us itself like for education sector. I believe that since we are so slow in adopting new technology By us will take care of itself outside the education in due course and and let AI Help us in improving our education. That's my viewpoint. Excellent. Thank you very much I don't have a lot to add to this When I saw the question it was like does anyone right now Know how to address the potential biases. I mean, this is a very new field. I think the whole world is grappling with How to address this and any any early adopter sort of has an ethical duty to consider this and I Think I'm I mean it's hardening that very early in the kind of discourse about this This has been picked up and is considered important Our institutions Equipped to address this well my bias here is that I work for a university and the kind of discourse I see in my university About the potentials of this new technology. It is a culture that worries very much about biases and We are certainly aware looking at our data that the out cut, you know Our outcomes are not equitable despite what we would like to be the case I'm sort of heartened that actually what Elizabeth was saying that By distilling a lot of data. I mean a bike bike kind of bringing out these biases That were previously perhaps hidden I think AI might actually move the whole bias kind of I mean We like to think we live in a West, you know in an enlightened culture But these biases just persist decade after decade, which is really rather depressing Here is another opportunity another kind of point of being given to us about quite how biases we are without really meaning to be and You know as a society Yeah, so by being a sort of mirror of us, maybe it helps us see ourselves and Feedback is how we learn maybe this is a form of feedback that we still have a lot to learn here, but I Mean I feel I would trust the average university more. I mean Big internet retailers will be using AI to sell you more stuff They probably mostly only care about the bottom line I would trust the average university to how take a slightly more enlightened approach Whether it is I think no one's really capable of this yet But I would trust the average university more than the average kind of commercial people But that's but also no one knows right now So let's all work this out together and try and make our own contributions to pushing the the sort of world in the right direction Okay, just maybe rooting on what others already said maybe I could add that Yeah, some scholars argue that We should have like a human rights-based approach to AI in education meaning that AI system systems Should be both designed and used inclusively And they should include this kind of human rights safeguards And in principle that sounds like a good idea, right? Of course, but I don't know I'm not I'm the very skeptical of a technological solution to this at least in the near future As long as people are uncomfortable talking about bias will suffer from it and our I think our technological tools will mirror this reluctance but One practical thing we can do is to address the issue Just like a Maya and a bit Arte and Sarah are honoured in their excellent talk on gender bias on day one of the conference So we should dress and demonstrate AI by biasing the classroom as well Just as forward thinking educators already talk about various forms of bias and use that as a starting point For critical engagement with the subject matter, whatever it is Thank you, that's really really really appreciate the responses So we'll move on to the second question And if anyone can pick the pattern here we now have up and I have an image that I generated AI I am and let me try that again AI generated image of our partner here Logislate framework so currently EU's proposed legislation and there's also legislation other parts of the world around regulating AI In some cases the AI may be considered use of AI making to the high risk in certain regulatory environments So what do you think the impact of those sort of regulations might be on the education industry? Rajesh, would you like to start on this one? Sure I think that picture actually is very relevant to AI in the sense that AI is so Potentially powerful that if not regulated It can be quite destructive. So I think More or less we all agree that there is a need for regulation Now what will be its impact or education industry? I believe that in short term There will be minimal impact mainly because as I was alluding to earlier is that The education sector will be slow in adopting AI so by the time it gets here it will be a well-regulated in terms of bias and risks In longer term, I believe that it is certainly huge risk in the sense that if we start using AI for personalized learning and Tutoring and that tutor is so biased that it can brainwash the kids at early age At a scale that is scary because now you can have a tool which can Generate content which is very biased which can really if the purpose or Ethics of the AI in the beginning is set to Really build a next generation of kids. We just follow the government Follow the government and then we can see the risk where it is going. So in terms of long-term, it is very difficult It very risky in short term I believe one thing with where I like to bring up is the content creation where Since so much of creative content is being consumed by AI without compensating the creators and and that's very very relevant for Education institutions. So I believe a regulation there will be very very helpful. I think I mean with regulation Obviously, it's necessary. So we do we need checks and balances I think it just really goes down to we're educated people, right? We know that we need to Be responsible with these things. And so While it's necessary, I think we do just need to have like proper things in order and understand the ethics of What needs to be regulated in my kind of lens, right? I'm in health care. So Looking at how we're gonna be utilizing this we want to keep people safe And I think a lot of us we need to understand human behavior, you know, what's most advantageous, you know Everybody is going to want to utilize AI to their best benefit That's perfectly fine, but is that gonna go too far? So we do need to make sure that we're keeping people safe and This is a brilliant tool that we can exercise greater limits of our creative and cognitive abilities. I do believe But we should be wise And just making sure that there's still that creativity and we can utilize it to its full potential without Anything breaking ethics, I guess We can't wait for the regulation We need to be the ones driving it The people who in this room are people who are passionate about learning we know how humans learn We are passionate about technology. We are the however little you think you know about AI right now You know more Then most of the people that you will walk past on your way out of the the conference center today so What I think about this is that in the education sector whether we're talking about higher ed K through 12 corporate education NGOs we should be the ones exploring this and driving what this legislation should be Because we are the most informed about it We need to be the community that starts talking about this and talking with each other and talking with our students Who whether we wait or not are going to be out there doing things with AI? We need to be leading the legislative effort Yes Well put I Think yeah for me As I said before we don't you know This is very early days in the world of super powerful AI and I'm reminded of a line that Martin Already used in his keynote that with great power comes great responsibility Yeah so From what we can see now. This is extremely powerful technology Extremely powerful things do need to be regulated but It is I mean writing regulation is not easy right if it was easy you just have one law that's be good That doesn't work unfortunately So Are we in a position do we understand the space well enough to write good legislation? Because even pretty good legislation I mean GDPR was written with a fairly good in understanding of how the internet worked and Is as legislation goes generally pretty good, but it still has the unintended consequences of Every website you go to has an annoying cookie banner. Is that actually making the world a better place? So I'm a little bit worried that Pre-emptive attempts to write regulation while they might Seem desirable Might have very bad unintended consequences. I think there's a bias here again Large powerful players in the world who are probably the people who most need to be regulated Probably have the kind of legal teams who can negotiate Big complex legislation What legislation might do is kill off the ability of small less powerful people to do Interesting research and development and exploration Without really stopping the big guys So that's my worry here that it disproportionately Impacts the people who could do the most interesting Work, so I'm sure this this I mean this needs to be done ethically I'm not sure we yet in a place where legislation will lead to the kind of regulation. We actually required and at this moment really While there might be risks the world needs to explore this and work out how You know find the most ethical and beneficial ways to use it and And see how it's abused and then once you know how it's abused then you can write more targeted legislation But yeah, so that's my my fear. It's difficult to legislate yet We will need legislation, but also we need to have the freedom to explore this ethically Yeah Yeah, so the the EU approach to AI legislation equates trustworthiness with Acceptability of risks and and some scholars have criticized this approach to AI Where the central problem is seen to be AI's trustworthiness? for example, I think it's reasonable to say that an AI might be trustworthy but still exploitative in some other way for example and Yeah regarding what team just said that so in the EU draft AI related education Educational AI systems are considered to be high-risk AI in the EU's classification I need to be assessed and registered in a database if the current draft bill will pass and It's gonna be interesting to see will it will it just like prevent small actors from playing or what what will the practical impact be but So I have no clue, but actually I was sitting yesterday in the panel. There's somebody here who works In the EU administration, so I would love to hear their thoughts if we have time after the round Maybe they have some insight Inside the information on this one Excellent, thank you So the last question and this is the fun one in ideal word AI has the potential to have a huge impact on education A number of fronts. How do we think AI will have the greatest positive impact to education? Yeah, and I'd really like the Panel to think to address what what areas or what uses do you think that you would what what things excites you about AI particularly? So Tim, would you like to start on this one? Yes, I Have no clue. I Think it's like asking someone in about 1993 what will the biggest impact of the internet be? And would anyone then have predicted let's say Facebook? You know, I think so I think really and You know come back and ask me in ten years, and I'll give you a good answer Actually, I think the biggest impact is going to be something we really cannot even imagine now and quite a lot of us will be surprised and That's sort of quite exciting. I think So I think it will change our perspective in ways that meet. Yeah But again, I'm quoting Martin, but I think we're never going to lose Educators For me, I mean it comes down to what is your definition of education? But it's for me. It is one generation Passing on its knowledge and more to the next generation and that is always about one group of humans bringing you know upskilling the next generation and Powerful technology is only going to let us do that better. We all know in this room, you know again going the parallels of history how the internet has helped us Bring more education to more people and again as a powerful technology This can only help us bring more education to more people But as a tool to augment and scale up what what one teacher can do Because that's always what technology has done. It's made it's made things that were previously exclusive You know Going back way into history in like in the 18th century You were either rich enough to pay a servant or you did your own laundry And now we all have washing machines, which seems a bit of a mundane example But you know technology has enabled more people to have stuff that was previously exclusive. So Again can't really predict how but I think AI has huge potential to do that sort of thing We made you honest to go Brett. Oh, we would like to speak next Yeah, thank you Yeah, well like like Martin outlined in his keynote I think we need our best bets are openness and privacy-minded and I suppose most of you will agree with that but Continuing on what Tim said regarding like predicting AI's impact I think it makes sense to look for answers from the history of new media adoption For example, the media historian Lisa Gittelman has observed that When a new media technology emerges, there are there are always lots of expectations of what kinds of changes will take place But generally speaking these predictions will be of the mark and Obvious point of comparison is the birth of birth of internet with the creation of ARPANET But I if you will indulge me I'll Make go on a small detour and talk about Thomas Edison and the birth of the phonograph tinfoil phonograph So it was very perplexing to people in the beginning And not that different from the people's astonishment of using chat GPT during the past year So Edison's tinfoil recordings. They were actually inscribed on tinfoil paper There were a but they were a black box of sorts because it was impossible to decipher by humans like How did how did it work for why by looking at it? And it was working against the expectations of people in Discriminately recording everything and that was that was astonishing about it And any Edison thought that this this would be used for recording, you know the like lofty things like speeches of the presidents for the next generation or something like that But people enjoyed hearing like fart sounds played back on the device or stuff like that sounds sound familiar with internet So the same thing with ARPANET. They thought that it's gonna it's a it's this new handy way to pass RFCs between institutions. That's Really great, right? But then the actual internet happened. Yeah, so So We simply don't know what the impact will be and it might be much more banal than we Expected even though just like the phonograph phonograph on the internet. They were on unbelievably useful for education in the end But what do we know again based on previous research is that when new media is introduced in an organization the existing genres of communication within that organization will change for example, let me give you One more thing about concrete concrete example on education reflection papers When you send chat GPT some lecture notes does a decent job at converting them to reflection paper I'm sure many of you have tried that But when you ask it to add critical reflection to the first draft, it's actually pretty good It can make nice critical reflection to the reflection paper. So I was astonished how That critical reflection is actually not a faculty solely of humans But it's actually an educational genre something that GPT can produce quite fluently so I think the greatest positive impact will come to those who are willing to accept this this These kinds of genre changes in education and modify their style of teaching But as as Fred Dixon argued in his presentation on Tuesday what learning means will still say stay the same You know if one wants to learn at least a little bit of struggle is required Okay, so There are so many possible impacts and you know Tim is right. We we've just got no idea. We're in the rocket ship speeding We don't even know what planet we're headed towards but Let's talk about a couple of positive impacts that are I think related so you know We've heard about the idea that Chat GPT or another large language model like that can take a bunch of notes write a reflective paper even do a Critical reflection in that paper and then we're all thinking okay So then the student just hands that in and that's the end of it But what if the student uses that and reads it and says oh? Yeah, but I would change this bit and Now we're talking about using AI as a learning scaffold So, you know, we talked about earlier a couple of people this this week have talked about Vygotsky and zone of proximal development big long complicated word But it basically means you want to try to have learning at a point Learning takes place best at a point when the task is hard But not too hard and there are a lot of things that new learners try to do and when they first start They're too hard So one of the things that I think could be an advantage For AI is to give people the ability to start to do something that they thought was too hard So I'm gonna finish this with a personal anecdote. I have all these ideas in my head people who know me are like, you know, I Could like explode like an egg in the microwave. I've just got ideas coming out of me all the time But I can't draw my handshake all the time not just when I'm up in front of an audience and I Have ideas that I want to Get out there so that other people can see them and some people would interpret them best visually and In my head, they're all in glorious 3d moving over time and I can't get them out there We're well I even had this little card game that I wanted to work on that would require an illustration on every page And I tried to hire illustrators and they would never get back to me And I tried to use clip art and it kind of looked chunky and I use 3d software to set up my scenes And that worked but it took forever and I did all of the illustrations for my card game almost in one night playing with mid-journey It gave me a voice for this idea that I had inside me that I wanted to share with other people Will anybody else care about it? I don't know yet Will it be you know about as interesting as Rainbows in unicorns studying at a university which is charming for about I don't know 30 seconds and then we're done with it I don't know. I don't know if anything that I did is Worth all that much but a couple of the things that I came up with I really liked and I felt like they gave a voice and Tim wants to react to that so I'm going to give it back to him Well again, sorry harking back to Martin's keynote, but Martin said when he goes out into the world what he wants is From other people from from content is a human beings voice and You know point of view and what you're saying is this technology has empowered you to better get your point of view out there Which is kind of great I mean, that's which is what we want. We want to allow, you know Empowering people to express themselves better is a fantastic use of AI So I just made that connection and wanted to highlight it The only thing that I'm going to wrap up with and then I'll pass this on to the other great people on our panel is People have talked about well, is this the end of creativity? No So again, I'm going to inform you of research in the field A paper that came out a couple of months ago showed that if you train a large language Model like chat GPT on content that was generated by other large language models the quality plummets You need human beings to feed the AI at least at this point and I think it's going to be true for a long time You need original content going in and ideally we should be compensating the people who generated that content But in in the real world we all do learn from each other and borrow ideas all the time How many times if we quoted Martin up here? So without the martins of the world feeding into these systems the systems are not going to produce anything very interesting and The other thing is that how many images did you have to generate before you get one? That's worth sharing the AI generates a lot, but as humans. We are the curators We are the editors We are the ones who who sift through that and find something interesting that speaks to us that helps us speak to someone else So that's our role. There's still a very strong human role in this and that will be the human role in education to help our students Sift through right good prompts recognize something that's worthwhile and be part of Sharing those ideas those human ideas back into the system and making them so that other people can find them and react to them Wait Tim just reacted to me and I just reacted to him This is education happening right here And was looking so that took me about 30 images to get that one, so I'm gonna totally play off that because I think you know what I would encourage you guys always look at this picture What does it make you think of like go back to second grade? What was your dream job and when you were playing with your kids on the like other kids on the playground? What were you doing making up stories and storytelling one of the things that I always and In sharing hope right is telling my clients your story matters I know in allopathic medicine what I felt really robbed of was hope because I didn't know like there was a future It's it's almost like because we get so analytical sometimes and we're all learners. I know we can all get in those modes Sometimes I feel like it kind of stops our creativity, but I think with this what's beautiful is We have an opportunity now that if if you take away our responsibilities like if you guys didn't have to earn an income You know, it's a very like cliche statement, but if you didn't need money, what would you do? We would create we were What in your in your day job you help create you help learn you help expand you grow right? That's what we were created to do like God it created with a plan and a purpose and to carry out our individual missions And we all together elevate each other in different ways And what I think AI is really going to contribute is Man just the opportunity to dream again. We need more Martins. We need more visionaries We need operators to so maybe we can bridge those gaps Maybe AI can be those operators in our lives those integrators And it's gonna allow us to do what humans do best and I really believe that is to create so I'm very excited about it. I Could probably talk today's No, it's it's true that AI is going to unleash our creativity to and and in an original level and I don't know how the future will be but if I have I'm given lots of resources to build something which will have biggest impact on education Then I like to do two things. So one is in short term. I like to help our Teachers to be more effective in giving personalized learning to the kids So my my son is like in grade six And I have seen it's a very small classroom just 12 kids, but still I have seen it is very hard to know which strength is at which level in the same classroom and remembering that in Every session so even if we know that okay, this kid needs help in this concept But as a teacher we forget and it is hard to remember and juggle between tennis to and so he needs help here that Seen yourself there. So having an AI tool and assistant to the instructors who can really understand the students data and then help the instructors saying okay focus on Explaining this concept for this kid or give an exercise of difficulty level at this so that There's a good chance. He or she will succeed and have more confidence and move ahead So I like to build that AI assistant for the instructors. That's the first thing and Second thing in long term. I like to I really enjoy virtual reality and I believe it has a really huge potential to education impact but creating content which is engaging and interesting in that virtual reality for as simple as creating an experience for Say using prism for light to come in and be able to see different colors of the light, right? It is going to require lots of effort to build that and I think AI can have a huge As you say like we can all create those experiences for our kids and for learners and put them in virtual reality and that will expedite and Personalize the learning in the way, which I believe will have really good benefit in the future. Yeah Thank you. Thank you so much. I'd like to I'd like everyone to thank our panelists. That was amazing We're very close to time We're actually running a couple of minutes over because this is such an important topic and we've got such a good Insights from the panel here when we'll take one or two questions and these are open questions So if anyone do anyone anything they want to ask the panel on AI Sorry, can we get the mic down here please we've got two questions here just very quickly one thing I haven't heard much of about AI this this week is the huge power electricity requirements for powering these things. I mean a Google search Versus a you know Google Bing search. I understand it's you know multiples of power needed to power the You know one versus the other I'm not sure I've a lot of question on that But I just wanted if that might be addressed at some point. Would any of the panels like to come it? This keeps me up at night It even when I told the story about generating all those images I was very conscious of the amount of energy That I was using up to satisfy my need to try to get these images out of my head and into a form I could show to somebody else Algorithms are getting more efficient We are trying to work on having better ways to generate the power Every time we take Something that humans do and try to automate it. It uses energy. This is limiting factor Ultimately We use a lot of energy for a lot of things I personally would rather have the AI generate a few images for me then keep my house five degrees cooler in the summer But we do need to be mindful of this that is a really good point. Thank you for bringing it up Yeah, thanks for bringing up the problem of sustainability and it's not just the power requirements I recommend looking up the work of Kate Crawford. She has this book called the Atlas of AI that documents all this The ways in which these big AI products are unsustainable And for example, there's a lot of recent work on What kind of new Chains of externalized rape labor is required by the fine-tuning of these models We like to imagine that there's not just like magically work, but actually it requires like huge amounts of manual labor Done in low-income countries often exploitative job conditions. So it's just not just the power. It's the resources It's the exploitation. There's lots of problems with AI on that on the social side Thank you. Sorry. We had another question here. Yes I think everybody in this room by virtue of being here is fairly privileged and established in their career. I Wonder how we can prevent AI effectively kicking away the ladder from people that perhaps You know doing the more junior kind of tasks that a lot of us went through It feels like a lot of when we talk about AI being like another team member Is it actually might be very useful for us at the stage we are in our career? But how do we ensure that people that may be coming up through university and so on and getting to the stage we're at Aunt You know held back by AI taking away everything they would have once done Are you I just want to understand the question. Are you saying that like it's a different learning path that let they should be They should go through the same kind of learning that you did or I mean a good example is is computer programming You can do a huge amount. You can get an AI to do a huge amount of programming now So as an experienced developer it can do a huge amount for you If you're learning if you were junior developer, how do you actually learn is it is it? Taking away opportunities to learn because it's actually just doing it for you My perspective is no, I think that it just changes it Humans are extremely resourceful. So I think if we need to learn something in a situation for like coding I'm not a developer. I'm a practitioner, but I know enough to be dangerous So I've had to use some like HTML for example because I wanted on my website to do something and I was impatient So I figured it out. I'm not a developer, but I was resourceful. So in that situation I think that AI can increase again our critical thinking abilities and utilizing because every single person in here is Equipped to do really whatever you want to do you make a decision every day So I think it on the contrary it actually expand our ability to learn more and more capacities in ways that maybe we would have limited ourselves And I just think it's going to change learning. We have a different bar Whereas you know when people went to computers and in colleges I'm sure they had the same questions of like, oh, we did math by hand. What about a calculator? Is that cheating? Like no, it's just going to change like we're just gonna have to be aware of that Sorry, just very quickly on a very small I am going to the calculator My father was when he did his engineering degree is the first years they used to calculate a university in South Wales He hated introducing our calculators because the problems became much much harder. So Well, I just think the lesson from history is that this is going to cause major sort of seismic shifts in the social order I mean and And How the human race manages that is going to say a lot about us, but if you look that, you know industrial revolution Destroyed a lot of what were previously manual jobs and The world had to adjust and find other jobs In every revolution has put a lot of people out to work finally for kind of the people in this room Middle-class educated coming for us now. It's not coming for the kind of manual workers. So One thing that's a bit negative to be a bit more positive along the way And you know in the past it was a kind of six-day working week if you were lucky It's now more normally a five-day working week So by and you know, so one effect of requiring less manual labor is The general standard of living has risen and the amount of time we spent working has gone down For most of us So again, hopefully this can ultimately after a lot of painful sort of social seismic social upheaval that hopefully we can Manage better than in the past Hopefully the world will be a better place where we have more leisure for ourselves and less need to work, but I I am worried about how we will get there Let's all do our best. I just say to like on that is even if it removes like some jobs again Resourcefulness if you guys lost your jobs tomorrow, you need to eat You'll figure out another job and you might who knows even enjoy it more So I think a lot of times we're like, oh, no, there's change and there's uncertainty And it's like human behavior to say that's scary But faith over fear guys, you know, like we'll figure it out. We will get through it I think that the more open we can be to just embrace more opportunities. I mean, I'm toxically positive But that's that's my perspective. Sorry on that on that point actually another aspect is In the past what proportion of the human race were artists and creative as a living My guess is that has gone up dramatically during the 20th century, but I'd love it if anyone's got the actual data So again, that's kind of uplifting the human race away from manual labor for sort of subsistence and survival and towards a world where more people can explore what it means to be human It will be up to us as the community To make sure that we don't leave anybody behind that that nobody is going hungry That people have still have the opportunity to contribute constructively and to feel like they have meaningful work and a meaningful life The jobs themselves will change It's up to us as humans to make sure that we're still humans and we're still taking care of each other Thank you. So we are well and truly over time. Thank you so much for the intelligence questions Thanks for everyone's attention and once more. Thank you