The Good News Episode

He woke up this morning and the first thing he did was get himself a glass of water. He didn't ask anyone, didn't wait for a carer, didn't ring a bell, he just did it. Later he closed the blinds, pulled them shut against the afternoon sun on his own for the first time in years. None of this should be remarkable, except the man I'm describing can't move his arms, can't move his legs, can't speak. A stroke took all of it in an instant more than 20 years ago. So how does a man with cragipelea fetch his own glass of water? Not a miracle, and not one of those gleaming humanoid robots you've been seeing doing backflips on the internet. The ones with faces and fingers and a billion dollars of hype behind them. The thing that gave this man his independence back has no face, no legs, no drama. The industry barely calls it a robot at all. It's an arm on wheels, and that's it. I'm Stephen, and this is the AI Transition. And with me as always, to keep me right, is Lauren! Stephen Howey, I'm so excited about this episode. We needed a lift after our last one. This one's got uh some heartwarming stuff. It is. This is the good news episode, right? And we are going to keep this one positive for the next 30 minutes. We are, and it's look, it's it's you know, focused on something close to home for all of us. Health. There's so many, you know, uh blazers being trowed and a few other areas as well. So I'm excited to get into it. We've got some really good and really interesting and heartwarming and profound stories and news items that are happening that are making a real tangible difference to people's lives. This is the AI that maybe was part of the you know the hype a year or two ago of it's going to be curing this and helping with this, but has been lost somewhat in the the doomladen narrative that has taken over, especially in the last couple of months, right? So we're gonna focus on these positive human aspects today. Yeah, and that's the good news. There is a lot of positive news out there. And you know, the similar technology that we're hearing, those really, you know, dark headlines about being used for good. So yeah, let's get into it, Stephen. Let's let's jump in. So let's do health first, because we've got a number of health stories. Um, do you want to give the first one, which I think has come from over in the UK? Sure, and this is probably close to home for all of us, anyone who's had themselves or a close family member going through the process of, you know, being diagnosed. Um, and or even with cancer major illnesses, going to the the GP and trying to understand what's going on with them, these seemingly innocuous kind of symptoms that you discover. Yeah, we've got some leaps ahead in the the GP's office. So one of them being, I think it's uh look, 8% doesn't sound that big, but when you think about the volume of people going to their GPs, and this one's come from the UK, but I don't know about you recently, um, Stephen, but going to the GP more and more, they're using AI to actually even just take notes, but all the same, yes. Yeah. In the UK they trialled um a new tool called See the Science. Good fantastic name. It's used in about, I think, 1,400 GPs across England. This is about 15% of all the practices across the U UK trialed this tool. Um and it demonstrated that, you know, the population was covering about 420,000 people. That, you know, the average GP without this technology probably detects about eight cancer cases a year, you know, where you've got a patient needing to come and see their doctor about three to five times before there's a risk or they start to pick up, you know, some kind of correlation. So and we all know how horrible this illness is, but by using AI and actually being able to correlate across the whole history of this patient, they're able to pick up uh the risks of cancer and send these patients off for tests early on. Yeah. I and and this one does hit home for me quite personally. So I've known a number of people uh recently who got caught early and also sadly some who didn't. Uh and when you say that 8% isn't necessarily a lot, you know, 8% when you're talking of 420,000 patients there, you know, or or you're talking about the tens of millions across across a whole population, 8% actually that's big then. So we'll be able to do that. When you yeah, when you add those numbers up, you're right. It's not inconsequential. And you know, a lot of these uh and look, we to say that the death word, it's obviously, you know, very dark, but two-thirds of these cancer deaths, right? They come from these non-screenable cancers where you go to see your GP, can you I can't even imagine like the amount of context shifting and 10-minute appointments they've had across the day to try and dig deeply into your history to pick up on these nuances. So now you might have a patient, I think the example was presenting with, you know, I've got a bit of a funny, uh, fuddy guts, as they might say in the country, a funny tummy, and uh, you know, a bit of diarrhea. And you know, usually a doctor would be like, oh look, if you're not feeling better in a couple of weeks, come back. But in the the case study here, we had AI running through that history and going, hang on a second, this doesn't look right. Let's go and send them off um uh for some diagnoses and actually finding, like you said, these cancers early. And and and that early detection, especially in cancer, as we know, is absolutely crucial, right? So that early intervention, you can actually get treated and have kind of full recovery and go on and live a normal life. But without that early detection, um then yes, you might not be um around that long. And even, you know, personally, I'm of an age now where I've had my here we go with the TMI, I've had my first mammogram, right? And even that process of going through your mammogram, it takes two humans to actually analyze the results of your mammogram. So you've got to wait a couple of weeks for that first great um review to come back. They've now, again, in the another trial in the UK, using AI as a second reader. So where you usually need two specialists to read these mammograms, they're using AI to sweep through and highlight, okay, well, here's some interesting-looking cells, and then getting, you know, the human element to come in and review that. So the detection rate, it's jumped from, you know, 7.5 to 9.3 per thousand women. So this is again finding the the early stage cancers early on. And you can imagine the human fatigue of reading all these scans as well. So we're not eliminating the humans, we're bringing them to the fore after we've actually used that the AI to detect it. And I think as part of that study that we were looking at, it was exciting that there's a already a 29% shortfall of clinical radiologists in the UK. So if you've already not got enough of the radiologists, having this as a workforce multiplier, it's not a replacement. I remember even a year or two ago, people were saying, you know it's pointless going into radiology or any of that sort of work now as a doctor, it's all going to be replaced and it's all done. It's actually not working out that way. Because as you said, this is doing the initial screening that's there. It means that you're able to look over a far larger population, you're going to catch a lot more. And this study that we're we were citing there was you know a breast cancer screening study. That was 175,000 women who took place in it over in the UK. Um, and you know, as you said, that detection rate, you know, went up from this one was 7.5 to 9.3. That's but there there's you know an extra couple of lives per thousand, you know, times that by you know, up you know, 175,000 just in that study. This is significant. This this is actually affecting people's lives in such a positive manner. I think also any, you know, and sadly, so many of you know, my partner, family, friends have had to go and get biopsies. And if anyone's ever had a biopsy, it's incredibly painful. And it's because you can find so many. Oh, this might be a, you know, we're concerned about this, let's go and take a biopsy. So look, obviously not a caveat, obviously not a medical specialist, but this gives you that hope that maybe we're we're finding these earlier and we're not having to put people through so much, you know, trauma and waiting weeks and weeks and weeks to find out what's going on. So again, you know, it was so um pleasing to see this that again, we've got this technology helping out in this health space, which we kind of hoped for. Because there's there's there's there's so many of these ones. This one is uh epilepsy and being able to diagnose that early, and particularly in children. It was a research that's been done here in Melbourne in collaboration with the Murdoch Children's Research Institute. Yeah, this was incredible. I think you you found this one again in terms of, you know, demonstrating the the ability for these tools to detect really minute lesions with up to 94% accuracy. So uh it just it just blows my mind that, you know, humans have been able to detect these before and to actually give them this support um so that it could actually lead to better outcomes. So some of these children with this severe epilepsy, um, you know, they've been able to lead a seizure-free life because of these surgical outcomes. And again, with these micro areas of lesions that these, you know, doctors and surgeons are trying to find and target in these surgeries that often are drug resistant for kids as well. So there's a lot that obviously goes in behind this in terms of the amount of scans, the MRIs, the PET scans, they just produce so much data. And you imagine humans again, um, not that we're taking them out, this to support the human experts, right? Um, trying to actually scan through these micro layers. And after a while, your brain starts to play some tricks on you. So the AI can actually help you spot these with up to 94% accuracy and then allow the specialist to take this further. So I think there was a trial here where, you know, there were 12 of 17 children went into target targeted surgery and 11 came out completely seizure-free. And this is game-changing for them. So these are seizures that can are completely debilitating for for young children in particular, where you don't know if that's going to happen when they're riding a bike or when they're swimming. And you know, it just it affects every aspect of the life. These children were able to have this surgery that then completely transformed their childhood and the rest of their life. I think the analogy was that they were trying to find the needle in the haystack of where this was, but the needle was actually a a piece of hay itself, right? I mean, it's it's that right, it's that tiny that they were trying to find. But AI was able to do that and then pinpoint and then let the surgeons go in in order to transform these lives. Huge. And again, this is that multiplier, it's not a replacement. It's helping, you know, our experts actually find these easier. So, ah consider the heart wormed, Stephen. This is amazing. So we've we've we've got some um, you know, case studies here on breast cancer, on all sorts of cancers, really, epilepsy. And there's also, and this has probably been a around for a little while in the mental health space as well. Yeah. I and look that this one where you do have to be careful with this because there's a lot of things popping up. You you go on to the to the app store just now, and there's so many AI coaches that are out there that are just basic front ends into, you know, kind of CLOD or Open AI, et cetera. And you do have to be really careful what it is that you're signing up to. But this one in particular was uh a landmark study that was done by King's College in London. It was looking at the mental health wait list, and they they created an AI agent that was actually properly trained on three-quarters of a million hours of type therapy and using CBT, so cognitive behavior therapy, you know, alongside human support. This one had some dramatically positive uh impacts. And again, huge numbers here. This is like, you know, uh, we're talking in the UK, 1.2 million people on wait lists to, you know, receive some help. It these are, you know, obviously hard times, stressful times that we're living in. And there was actually really good results. Like I'm always, you know, probably like yourself, a little bit suspicious of like, oh, how do you feel if you're getting therapy from a bot? But there was a lot of comfort in knowing that actually there was a human nearby that they could actually tap into if they needed to. And they saw some, you know, really incredible results saying that there was, I think 80 odd percent showed a clinically meaningful decrease in anxiety symptoms. So they've got this app that they can work with. So imagine in the in the dead of the night when your brain is just racing all the you know the torments that it wants to take you through at night. Um, they can actually, you know, access this app. It'll take them through CBT techniques that it's learnt over the course of you know 750,000 hours and actually have a real responsive conversation. So it's not a video or worksheet. This is a live kind of responsive conversation they're having. Um and that human therapist is there, like we said, in the background. And I but I think as well you were saying, you know, in in the middle of the night, and that is one of the differentiators here that there is, I was gonna say someone, right? And huge quotes they're going on, but there's there's a there's an app, there's there's something there to talk to at two in the morning that is going to guide and help you. Whereas it's not like two o'clock in a on a Tuesday afternoon, you know that's when your next session is with the um with with the therapist. You know, you there is someone who's going to be there 24-7. Someone again. Be but you know what I mean, right? It's that access to that support. And maybe that's that moment in the dead of night when you're all alone that you can actually break the back of something and get you through to actually the next day. So again, we we've got to be careful. Obviously, there's some, you know, there can be some biases in these data sets. Um, but I think that again, that human element there that they can tap into. Um, and people have responded really well to it. And again, there's just so many people needing this help, and there's not enough therapists to go around. So And uh look, we said this was only gonna be good news episode, which is one little which is one little caveat. What did come up for me was the the irony that you know this might be needed even more because of the mental health explosion that's going to happen because of AI in the next couple of years. But at least AI is there to support you through it as well. So Yeah, look, we did we went uh but again, it it's a it's a trial, it's you're getting giving us positive insights where if it's used in that right way with the human pairing that you know, of all the horror stories we've heard, this has actually been a good news story. So where do we go to next, Stephen? Why don't we shift on to energy and particularly the grids, which we know quite well. You know, we we do you know work in that space um a lot ourselves in our in our professional capacity. We do and we're being grown-ups. Um but this is a really interesting study that was that happened using um AI optimized wind farms in the UK. And what they found was that um by using these these wind farms, and for those who don't know, that it might sound like it's just a kind of a standard windmill that's there, but actually, you know, they are all placed in you know specific locations, they are banked and you know, with different rows, they go in slightly different directions and kind of and pivot depending on what's going on. And how you do that, how you basically kind of just feather that positioning of it, can make a really big difference to the amount of electricity you're going to generate anyway. What they found was by using the AI systems who are doing this in real time and doing that kind of pivoting live and doing that constant monitoring, they were able to get between three and five percent more energy output. Now, that again might not sound like a lot, but actually when you're talking about you know uh wind farms and grid systems, three to five percent, that's huge, right? And the difference that has for both the energy that it's creating, but also maybe you need less you know, intrusive windmills that you know that people don't necessarily want in the one in that area, but so the ones that you do have or the ones that are there are optimized so you can get the best out of them. I think they were talking about no new turbines, but where they had the AI dynamically adjusting, and this is where you get to say some cool jargon, adjusting the turbine blade pitch in your pitch in your in real time. One million more homes powered without extra turbines, just because you've got that precision. Because you can imagine the complexity. Uh, and again, not to take away from the humans who are incredible, um, but this is where you can actually use AI to model this across time and get the the smaller nuances. Yeah, and when you're talking about humans being involved in that, especially with winds, you know, that is a 24 by 7 once. This is working at three in the morning, right? When the humans are largely in bed and you've got the monitoring that's going on. This for me is one of the kind of the no-brainers for AI where this is if this is a good, this this is one that is there's a positive aspect in here. Exactly. In the UK and everywhere, really. That's where you're trying to figure out how we use renewables and when we put the kettle on. I think there's a huge surge in the UK. Um, insert the name of a classic, you know, UK TV show where we turn all the kettles on and watch it at at night. And you've got that obviously the wind taking over at night, just one of the renewables, obviously. Um and because it is hard to forecast, right? You can't tell when the sun's gonna shine or the wind will blow. You can optimise this so much more dramatically and and improve that prediction window. No, I really like this one, which was um this might sound a little bit freaky for people, but this is use this is using drones to detect heat loss in certain parts of the UK, right? And then by doing that, and so basically what happened in the dead of night, you've got these drones that are flying over parts of the city. Not alarmed at all yet, Stephen. Not alarmed at all, right? And what they're doing is that they're looking for the heat signatures of the houses that are radiating the ambient heat that's that's going out that clearly don't have the insulation or you know the energy efficiency, etc., that's there. So basically what they were doing was seeing what was going on and then using that in order to target the households that needed the help. So rather than doing the blanket leafleting and trying to contact everyone in a whole area, etc., they were literally able to target those households to say, um maybe you could you'll be looking at you know heat loss prevention and this could be saving you thousands, etc. Made a huge difference. I I do struggle slightly with this because you've got drones morning through in the middle of the night about what's going on. However, however, um you heard anything over your shed in in Fairfield? But they you know and this was this actually won the Manchester Prize, uh, which was using these AI drones to thermal scan whole neighborhoods. Fascinating. Using those drones, using that AI monitoring, and then targeting, and then being able to get such a such a positive result uh that was coming out for for those neighbourhoods and for the people that were involved in it. Incredible. And there's even, you know, some uh smart green technology where you can turn your brickwork into radiators with these smart external panels. So you're detecting these heat signatures, there's just so much you can do. So, like you said, it's one of the actual good uses of a of a drone detecting these heat loss. I and let let's keep with the drones. Because uh big one here in Australia is obviously the solar farms that are here. Um and these these can be big, right? And so they're they're only getting bigger. You look at what's going on in the likes of China and actually places like Saudi Arabia, etc., as well, these solar farms are absolutely massive, right? Um I mean they are they are really, really big. But one of the enemies of these solar farms, especially as the years go by, is dusts, bird droppings, you're just kind of these solar panels getting covered and losing their efficiencies. And being able to have people who are going and checking that and cleaning that, that's quite the job. It's quite the job. So this is where you you tap your little drony friends and your AI robots, and they can actually obviously replace those weeks of manual inspections, and they can use their thermal cameras, LIDAR, sensors, they can get through the rough terrain, because remember, sometimes we've got sheep grazing under these things. Spook the sheep. So so so that one's not an AI one, but it's also one that I really love. So that there's a lot of solar farms now, and there's there's a couple here in Australia that the the panels are just slightly higher. Um, but what they do is they provide um shade and concentrated water runoff, which means that the sheep graze underneath the solar um panels, which actually saves a huge amount of money, so you don't need to kind of keep that registration down. So it protects the sheep, um, it's good for the environment, it's the there's it's like tick, tick, tick, tick, tick, tick, right? And you've got sheep grazing in these solar farms, which are now got these drones overheads cleaning the panels. Exactly, monitoring it and you know, flagging where again, not taking the humans out of it, they are flagging, hey, we need to go and fix the wiring of some of these panels, and they're building 3D maps and really optimizing the space. So uh yeah, again, it's quite incredible what's going on out there. Robot Guardians for Solar Farm, Steven. Where else have we seen some robots? I'm excited about this one. You kind of let us in. Um so look, this was this was more your connection here. So why don't you talk because you've got a you've got a personal connection to this one. I know. I look, uh I'm so it's uh chuffed and proud to be able to say I know this amazing person. Um shout out to uh the lovely Aaron, who has an incredible company called Hello Robot. And at the start in the cold open, Steven's actually talking about um an amazing human called Henry Evans. He's a nonverbal quadriplegic, and he actually sits on Hello Robot's board. He's been working with them um for over a decade now, and they've just released their new robot, Stretch 4. Um, and there's so much uh to this robot in terms of what it can actually do. We've seen a lot of humanoid robots out there, you know, your big scary human-like, you know, robots coming for you. Uh, Hello Robot is all about how we actually help our humans. So this isn't about creating a humanoid uh robot that can do backflips. This is actually about how we connect robots to the humans in need to serve them. So highly recommend jumping online and watching some videos around stretch because it looks distinctly different to any other robots you've seen before. Yeah. I know about you, Stephen. But it took me back to the eighties a little bit. Number five is alive, short circuit Yeah, it was a bit like that. Was that and Digging into this this whole story more, what what I really liked was um I I've been you know hearing for the last couple of years, if you listen to to Elon Musk, you know, talk about um you know why it has to be humanoid robots and why they need to design them to look like humans, etc. And you know, and the big argument there is that we've we've designed our our whole environment to be humanoid shaped, and so that's why you need a robot that that's humanoid shaped. But actually, if you're in an edge care environment, you actually have a quite different environment there where it's already been set up for um wider corridors, it's got it's got it's got handrails, it's got doors that you can get through. But also the the important thing there is all about the safety of the person that's there. The last thing that you need is one of these robots that has a bit of a malfunction and then starts tipping over because it's a humanoid robot. You want it to be if it short circuits, you know, going back to what you were saying. Yeah, look at that connection back there, yeah. You you want it just to stop where it is, and what and this is what it is. So for those who haven't seen the pictures of it, it's basically just a um a bit like a roomba, right for those vacuum cleaner things, right? So it's got that base big so so big, heavy, big, heavy, um basically where you put all the batteries, etc., with the big pole um coming up, um, which has then got the uh the vision and the arm, etc., on the top of it. Um brilliantly designed. Absolutely brilliantly designed. Incredible. And look, I won't, you know, uh we could go through a whole history of where Aaron's come from, but um digging through it to prepare for this, interestingly enough, he actually came from um he grew up on a farm. And do you remember, you know, you're a kid and you're like, right, how can I help, you know, mum and dad around the farm? And Aaron experimented with um actually building robots to help, you know, load hay bales. And do you remember the 80s you grew up, you know, and and sorry, kids, but you know, you actually had to get up and change the Pally, you know, Aaron actually created a robot um to go and change the television channel. So he look has an incredible background of working with some of the pre-eminent roboticists in the world here, but really it all came back to exactly what you said, Stephen. How do we help people um actually live their lives more independently? So Henry, who we touched on at the beginning, he uses stretch for really as an extension of his own body. So he uses it to, you know, get him some water um to close his blinds to play with his granddaughter, to brush his hair, like giving him a huge amount of independence that um before, you know, did not have at all. So Aaron works really closely with Henry to understand his needs, not to assume that we want a humanoid-like robot. So some of the tech in here is extraordinary. This these robots have 3D lighter. Um so that means that they've got laser sensors so they can see in the dark, they can make sure that they're not about to, you know, roll over your shihitsu in the middle of the night. There's incredible um sensors in terms of floor hazards. So this is all about safety and understanding what humans need, to the point that Henry's got some really um, you know, touching videos of him playing with his granddaughter. Before granddaughter was a little bit scared and couldn't connect, and now he's this incredible robot that can help him connect to his granddaughter. So they've gone from stretch three to stretch four, incredible leap ahead. Um, and these robots, again, he's not about going out there and mass marketing these robots. They really are about connecting to people that need them, people with disabilities. There's also some incredible footage of um some people that were taken to the local aquarium because the aquarium had hello robot connected to people at home who can take them on a tour around the aquarium, which I think you and I chatting before, um, you've your wonderful partner is in health and having, you know, starting to explore some of these technologies. And we immediately went, oh my God, imagine a hello robot in that um environment. So it's quite incredible when you start to think about that connection. What do the humans need? Not that we want a robot that looks like us wandering around our house because again, the house has been optimized for someone who's paralyzed. What can they actually do to help them get that confidence and independence? This one really hit home to me. You know, again, for me, this is just an unadulterated good. This is where things like AI should be getting used, the positive impact that it can have on people who, without that, are completely relying on care from humans, which is incredibly difficult to get, incredibly expensive. We know what's happening with the aged care systems these days and and and what's happening with aging demographics. Uh this is so important, this this sort of work. Uh, and it's wonderful to see um something like this that's going to make such a positive impact in actual people's lives. This is incredible. Stephen, have we managed to spend half an hour on this? It's been just so heartwarming. We're gonna put a bunch of links in the in the show notes so you can go and check this out. Yes. And the good news is there is more good news. We could just keep talking for quite a while, particularly as you said, in the aged care sector where we've got AI coming in to help people keep people independent and at home. You said we we've touched the the tip of the iceberg here, and you might have noticed that nearly all of this has been in the UK and Australia. It's not just happening in certain places in the States. This transition, this revolution, whatever we're going to call it, is it's happening globally, and there are positive aspects to this. I mean, we've talked about health, we've talked about the energy sector, and then these individual robots that are able to help you know, individual people who need that help. So this is so important, and it's lovely to be able to cover some stories like this. It is, and even, you know, look, I keep talking about my lovely friend Aaron here. Shout out to Charlotte, my wonderful bandmate who married Aaron, and I was lucky enough to, you know, go out for dinner with him recently and actually meet the person that's created this technology and see how much they connect through. And this is open source Halo Robot as well. This is for people to research on. So again, some really good news stories coming out of here. All right, I'm gonna stop waxing lyrical about my amazing, talented friends. Do usually we need uh we need a joke to kind of cleanse the palate somewhat, Lauren, but this is actually to kind of keep the fun stuff going. Do you have something that can keep us going before we wrap up? Look, I did. I really searched, I thought maybe I'll use one of my classics. Stephen, did you hear that ants can't catch the flu? Why can't ants catch the flu, Lauren? Got tiny antibodies. No, oh no. Classic familiar with my work. Yep. Classic Lauren bad dad joke. Love it. What a heartwarmer Stephen, thank you. That was lovely to do that this week. And look for those who have been listening in or watching in, if this has got you thinking, please share it with someone, wrestling with some of these questions themselves. Um, subscribe and comment. Uh, whatever you're getting us from, it really genuinely helps. And we will see you next time. And remember, you still matter, at least for just now. Thank you, Lauren. Thanks, David. See ya.