AI and climate changeAI and waste management
VR in healthcare
Data tracking
Prediction
Del denne Casen
Velkommen til Lørn.Tech – en læringsdugnad om teknologi og samfunn. Med Silvija Seres og venner.
Silvija Seres: Hello, and welcome to Lørn Applied AI. My name is Silvija Seres, and my guest today is Tom Allen, the founder of The AI Journal. Welcome, Tom.
Tom Allen: Great to be here. Thank you for having me. It's brilliant to be on the show.
Silvija: Being here means being in Birmingham, I guess?
Tom: Yes, it does. Nice and sunny in Birmingham at the moment.
Silvija: Yeah. It's funny how in these digital times, we've stopped asking people, how are you, and we start asking, where are you.
Tom: That’s how I start all my conversations.
Silvija: But at the same time, this is making it much easier for us to speak to people like you, which is a great advantage as well. So, Tom, I'm just going to say a couple of sentences introducing this session, and then we'll start our chat. So Lørn is a Norwegian, educational media company for corporates. But really, we want to educate the whole population, not just the working population, and we believe that the whole world needs to know and needs to learn. We want them to learn more about the new technologies, and how they change our business and our society. And at the moment, we're making a series on applied AI, where we believe that AI is a necessity, and probably the main tool of our futures. But at the same time, we know that way too many people are scared of it and believe that you have to have at least a master's degree in maths or computer science, to be able to talk about AI. What's most important to learn about AI at the moment is how it can be used as a tool. We've created some stories, and from my understanding, you're also going to help us expand that view.
Tom: I hope so.
Silvija: So, with that, I would like you to tell us a little bit about yourself, and then about your journal. And I'm always asking people to be personal in the way that they describe themselves. So, you know, what drives you? Who are you?
Tom: Yeah, love it, thank you. So, my name is Tom Allen, and I'm the founder of The AI Journal. And I started it around January 2020 and we launched in February 2020, and my background is in marketing. Very heavily involved with engagement and community growth and building the technical side from content development. Through working at an engineering company that led me to, and I knew nothing about AI and knew nothing around automation and all these technologies. That was in November 2019. So that led me into this whole new world of understanding, and it completely switched up my experience from being agency side to all of the sudden learning all about these new technologies. And through that, that kind of taught me quite a lot about who I am and what I've done. I didn’t progress very well at school myself. Through that, I kind of understood where to learn and for the university to pick up my kind of learning patterns. I think the reason I moved towards these technologies is that you have to see a lot and learn a lot, which is exactly how I learned compared to reading and writing. Through that, I got into learning and loving, which I wish I had learned at an earlier stage of reading a lot of books and reading and understanding different ways. That opened a lot of opportunities for me to develop the journal, develop my career and develop my roles in marketing. And yeah, I love to explore things and take big risks. I'm a very risk-indulgent person. And through that, not just in work and through chasing the opportunities with a journal or roles as in, but also in my personal life and doing skydiving and going all over the world and meeting new people. I mean, that's what I love. I got called by someone at Expert.AI a “blackbelt in network-marketing” because I just love to speak to people. I love to meet people and love to get out there and see what their story is, similar to what we're doing now.
Silvija: And now tell us about your journal. What's the origin story, and what's the reason for its’ existence.
Tom: Thank you. So, the origin story was that I started it in my bedroom as a blog. It was nothing more than that. And I thought it wouldn’t turn into much, and then the lockdown happened in the UK and we all got furloughed and I worked for it for what, six, eight, to ten weeks. Being an engineer, I had to go back pretty swiftly to look at automation lines and sales skyrocketed because everyone wanted to start using AI, which I'm sure we'll start talking about later. And their order values went up through the roof. But the journal has allowed me to use this platform to educate people. And that's what we're here. The core vision behind it is to educate as many people as possible on AI and emerging technologies. And a lot of people get quite scared or quite held back by wanting to learn about it, or it doesn't get explained properly. And through that, we found an opportunity to get, not just our side, so not just our reporters or own news team, but people from the business to come and talk around the different topics. And that led us to go into all sorts of areas, whether it's reports and webinars and reaching people. And it's been an explosive journey and a really fun journey. To give you an idea, we started in February 2020, now we reach around 220,000 people monthly. Our LinkedIn audience grew from 4000 in December 2020, to nearly 30,000 today, all organically, and it's great to see that we built this community. Nearly all our resources are free and it’s similar to what you're doing at Lørn.tech, to provide that value and to provide that education point. Know people don't have to feel panicked or worried about it but can come and learn in safety or in their own time. And learn from the best set of people, not just one-sided, but people whether it's the heads of AI at BMW, directors at Highland, or CTOs at Capgemini, they can come and learn from us and come and learn from them. So it's a great way to get the name out there and a great way to understand a technology that seems to be taken over every single day.
Silvija: Two things I love about this. So one is that it's democratizing knowledge. I think we all need to learn, not just people with PhDs who are invited to this future that you're all going into. And the other thing is that we need to learn from each other. I think that you know, the times where we all have to go seek a professor and drink from the source of knowledge are over because innovation happens, where it's being applied at the moment, and it's through these applied stories of innovation that I think we can all move as fast as possible. So, a wonderful concept Tom. And why AI? That's the other question I have. Why did you choose to focus on that?
Tom: Honestly, I don't know. It was one of those things that just happened and that's my honest answer. Because I moved from an agency background into automation. With automation, you have to do a lot with machine vision and a lot of automation lines and predicting faults and detecting patterns. It just naturally led me to understand what automation is, what machine vision is. I had to do a lot of nighttime reading, a lot of further reading. That's where it annoyed me, and I thought this is a topic that's exploding. I've seen all the reports. I've come across Gartner and Forrester and IDC and all these other reports from McKinsey and Deloitte and it just kept hitting. PwC especially I think they did a report that highlighted it's going to add 15 trillion to the economy by 2025. And it just hit me that this is going to be a big area, and I need to write about it. Because I don't know if it does exist, but I couldn't find any places that did talk about it from a ground-level view, from the operator's point of view. It was always journalists, and always editors who are editors and journalists, they're not technical people. Maybe they had a background in it, but they're not people that are in it every single day, which is why we wanted to democratize it. And AI just naturally became an area, and now it's exploding at a very rapid rate, and you're seeing massive funding rounds, or big acquisitions and big ideas coming through the door, so I guess luck, but also, I've never been that fully involved in tech, but because of my role I got launched into it very, very quickly as Head of Marketing at an engineering company and needed to learn about it. And it just to me seemed like a golden opportunity to talk about it and to help people because like we've already said it's hitting every single industry in every single way you can imagine. There's nothing I don't think it will be involved in by 2030. Only about three weeks ago, Sundar Pichai, the CEO of Google came out and said, “AI is going to be more valuable and more important than fire, water, and wind” or something, like basically the natural elements, which kind of gives you an idea. And Elon Musk to put a daunting view or a more optimistic view, said it's going to be more powerful and more capable of changing the world than something like nuclear warfare will be able to. So, it gives you an idea of how powerful this technology or capability is, and what it's going to help us achieve. But long story short I’d attribute it mainly to luck falling into it and getting in. I'm really glad I did because now we can provide a very valuable message and invaluable opinion to people.
Silvija: I'm smiling. I'm just re-reading The Hitchhiker's Guide to the Galaxy. Do you remember it?
Tom: I haven't read the book. I watched the film, but it's on my list. It's one of those ones you need to read
Silvija: So, there’s this guy who becomes the president of the galaxy and he doesn't necessarily know why, and when he starts thinking about it then he doesn't want to do it anymore. I think sometimes, instincts are really important and great. Listen, let's talk about applied AI. So, through your journal, you get to hear a lot of really good stories.
Tom: Yeah.
Silvija: And I believe in storytelling. That's what people remember. They feel like they're learning. They feel like being entertained while they're learning.
Tom: Yeah, 100%.
Silvija: So, let's just pick out some stories, some of your favorite stories. You mentioned to me that you have spoken to, for example, the CTO of Capgemini.
Tom: Yeah.
Silvija: What's their application of AI?
Tom: Yeah, so they touched on one area, which was in with climate change and waste management and how the kind of two intersect. I used to work at Capgemini or their subsidiary Capgemini Sogeti, which is towards testing and infrastructure testing, and security testing. But the idea behind that is to paint a picture of anyone wondering how climate change can help messes in Britain. We've written loads of articles on it, not just from that team, but from loads of other people. And the idea that it can help with weather patterns, it can help with predicting, and you've got to think about it as a full ecosystem, right. So, it can help with predicting planting foods or when the right time is, or how to predict areas of fires so wildfires that are happening. And these are areas that get talked about in the article and predicting waste management because if you have bad waste management systems, and it's all going into landfill sites that become overcrowded, you're going to have a natural problem where you've got a lot of waste in one area. So AI, and it's what I always look at AI say to people don't get it complicated. AI is essentially a prediction tool, that's if you want to break it down to what it is, it's a prediction tool, it can say what's going to happen by when very, very accurately, because machine learning will run thousands, if not millions of patterns. And through that you can understand where to plant foods, what the right time is to plant crops, so you get a better return on your yield, more money for the farmer so you can plan better, you can understand your cycles better. From a water perspective, you can understand where water use is going, obviously, with food and cows, the other day I read some stat about how many millions of liters of water it takes to feed one patch of cows, and also through that you can use it a lot more streamlined, and you can get water there a lot quicker. And through that, you can understand where the next wildfire is going to be. So, you can get the fireteam deployed quicker and stop it from ever happening, which is a bigger and bigger crisis, as we're seeing today with climate change.
Silvija: So, this is AI in agriculture and AI in climate support or management.
Tom: And you're going to see agriculture as part of climate change. People forget that because it's an ecosystem. If one part doesn't work, another part doesn't work. If you burn a forest down, you're not going to have natural wildlife. And that's how AI can help because it will say, if this happens, this chain of results is compound, that effect is going to hit this area and you'll be able to stop it before it happens.
Silvija: It can show you scenarios. I just wanted to give you one more example, a very Norwegian example, we did a series here in Lørn with a Norwegian innovation cluster called Seafood Innovation. They work with these fish farms in the ocean just outside the Norwegian coast. You know 200,000 to 500,000 fish in one pen. Now AI is being used to monitor their health and well-being, but also optimize the feeding patterns and water quality. It can detect not only a certain percentage of stressed fish or undernourished or overnourished fish, but it can detect an individual fish.
Tom: That’s incredible.
Silvija: It’s quite amazing by looking at the patterns, their gills, and so on. It's changing our world. So that was AI in agriculture. You also mentioned to me that you have examples of AI in engineering?
Tom: Yeah. So AI in engineering is a gray area because it's great for spotting detections and spotting faults, which is huge. I used to work in engineering, and I'd be on the automation lines and doing assessments and I wasn't doing any assessments, I was more looking for marketing opportunities, but I'd be with the engineers who were doing the assessments.
Silvija: Just a question, you're talking about automation lines, can you please give people a picture of what's an automation line?
Tom: Yeah. So, say, the clothes you bought from a local retailer, or you bought it from Amazon, Asos, or H&M. That's going to go through a packaging system, where it's going to have tons of automation. So, they're going to pick it out, arguably from a robot of some kind that will drag it through, it'll be picked off, it’ll be packaged. Through that, there will be a mixture of human and automation work and robotics. And through that, that will be packaged up, that will probably be folded very quickly and organized. That will go on to an automation line, that automation line will have a distribution service of where it's going to what part of the country or what other country it’s going to. That will use a line that no one touches, but it will automatically find its way to the back of a truck to take it up to your local depot, which then comes and then you get your delivery driver - here’s your new dress or whatever it is like a new shirt. And through that, it might be in multiple stacks, or it might be in single stacks. So, you can picture the problems that can arise from it. Because if it's all humans doing it and taking it off and moving, it builds up a time lag. With this, you can automate a whole system that makes it all through its’ long journey of a massive warehouse, and they're huge, they're massive sites, within five minutes, so it's no people running around. It's very systemized, very ordered. Through that, you improve your timelines massively because you've gone from point A to point E within five minutes. And that allows you to get your product quicker. Another example is wine. So, you can imagine the pain it would be to bottle each crate of wine that needs to go to a wine cellar. Or Coca Cola, another client we dealt with, and a robot will basically palletize all of that, spin it around, use a Stanley robot or something similar, which is for small goods, and pick and place it into the box very quickly. I'm talking like every 10 seconds or something. And through that, it will package it up, a massive gantry robot will palletize it, spin it around, plasticize it up. And then it will move across an automation line or conveyor system into the back of a truck. And imagine trying to get all people to do that, the weight, the load, the time, it just wouldn't work. You'd have a lot of delays, and you have a lot of problems. So hopefully that gives a good picture, especially with logistics.
Silvija: Right, so what AI is doing here is gathering production data. It has sensors in these robots. And it can find ways of making optimal paths through this factory, finding the optimal timings, and look for bottlenecks.
Tom: Exactly. So, you'd use something maybe like machine vision, because you have to capture all of it, or people are using more sensors, which is the “Internet of Things” or an IoT device. And that will be positioned, and you’re sitting at home at that moment, but not to get off off-topic. So, you'll see it within a factory or a warehouse, or a logistics place. And it will be monitoring every single move, every single fault, every single direction, and mistakes that happen with machines from time to time, did it go down the wrong path? And the good thing is if there's a problem, the AI, maybe using a camera or maybe using a sensor, will give feedback to the algorithm and it will learn that pattern so it doesn't happen again. If you try and teach a human that you're trying to teach it, God knows how many patterns that you'd have a problem with the number of people that you have within the workspace because you'd be all on top of each other. But a machine can be registering, if you stick a bird's eye view, it can be registering 50 different views all at once. It can be spotting how that parcel didn't go to the right returns area. Why that bet didn't go to the send-through area. And it will pick up all this data and it will create a feedback loop that just keeps saying “That's where that problem is. That's where that problem is” right down to the smallest detail. So, if you're going through, again, on an automation line, look at returns, if you're returning an item, they need to go somewhere and they need to go back to the seller or back to the warehouse, they’ll use machine vision to scan all the IDs using their RFID-scanning. And that will use AI to register where things are going. It can predict when the high times are, when do they need more staff in, what areas are getting the most returns, or what areas are getting sent out the most. All this allows you to predict better or give you better accuracy on where resources are needed and where things need to go. So very strong as a business because it can ultimately save you cost and save you a lot of money
Silvija: It Gives you a detailed insight into the organism of your production. Also, we talked a little bit about AI used in airspace for two different kinds of things. One is patterns of travel. And the other one is optimizing whether it's the airport or the airplane itself.
Tom: Yeah, you got two ways, you got the engineering side, and you can use something but it's not so much AI, but it's a simulation tool. This uses AI to a degree because anything that uses data is essentially probably AI. After all, it's an hour and a half to process the data. Through that, you can understand when the engine is going to run out when the breaks are going to happen, you can robustly test it. And digital twins are very good for that because they can test all, it’s a simulation tool, so it can simulate all flights or how quick the engines are going, or all the infrastructure for the plane to make it very safe. So, when you're going away with your kids on holiday, you know you're on a very secure plane that's been checked thoroughly. And it's not open to human error. So, you know, things aren’t going to break. On the commercial side, which is the fascinating side, is using the data, so when you go on holiday, or when your colleagues or anyone who goes on holiday, and you go to an airport, it's going to be registering, especially if you've got a smartwatch or a phone. They're going to try and track as much data as possible as they can to understand when do you travel? What time of the year do you travel? Where do you like to travel? Who do you travel with? How many people do you go with? Are you traveling for work? All these filters that they will start attributing to you to think okay, we want to give this person the best possible experience ever. We want to give them the best user experience which is, in an airport's situation, do they stop for food? Do they like to go grab a coffee? Do they like to rush through? Are they always last minute? And it will be counting all these things, maybe not over one trip. But over five years, they're going to get a very good profile of who you are, where you go, and why you go. And this is all AI. And what it will be able to do is create a feedback loop for you to say. You're going on holiday at this time, here's a special deal for this place that we know you love to go to. And there'll be all the follow-up details as well. We know you like this hotel, we know you like to travel with so and so we're going to send the same offer to her or him or to your partner or whoever it is. And through that you get a better experience because it's cheaper, they'll offer you better rates to get you included and to get you going. And it will be a better experience because they'll know who to target. They'll know where to go and you'll be on planes, or some people don't like flying with certain airlines. So, I'll be able to pick that up as well, especially from someone like Skyscanner who sends you these rates, maybe not the airlines themselves. You get the best rates and you fly with who you love. You don't have to spend as much time booking or searching. All those panics of where do we go on holiday, it's all suggested to you. And it's all that's all largely done through AI, AI, and data, but obviously, you can't do much with data unless you have AI. So, it will give you a much better experience and a much cheaper experience.
Silvija: Hmm, very interesting. So, what about smart cities? Maybe we can spend a couple of minutes on that as well.
Tom: Yeah, smart cities are the biggest area going in my opinion. It's an area that's going to change so rapidly at the moment because a lot of people are trying to build it in. And smart cities can be anything from traffic light systems to water and the infrastructure, how you get the water to your home. To cars that can recharge on the road, electric cars are becoming bigger and bigger. One of the big things is if you don't get to electricity points it’s a lot harder to recharge a battery pack than it is bringing a case of fuel with you on the side. And looking at all these different ways from smart motorways, detecting criminals, looking at people that are driving quicker so they can understand why are they driving quicker, where are they going? I think eventually all this data will be fed through autonomous vehicles. So, you have the optimum for your computer because that's essentially what Tesla has driven off its computer. That's why it's so valuable. Because the shape and design can change but it will be looking at how can I get you to your place the quickest time and it will be creating all these different interconnecting feedback loops, and not to get too complicated, but from a single traffic light, where they know it has heavy footfall. So, a lot of pedestrians are always there. Take somewhere like London, like right in the middle of Oxford circus. There's a lot of people, so there’s going to be a lot of people crossing the street, your computer's going to know that's not an optimal driving path. It will also know where water needs to go, how more water needs to be processed. And that goes into investment. So, it's a real, I don't want to get too broad, but it's a real hub of how do you give someone the best experience when they're traveling, or how do you make it more secure for them. And I think that's what we're facing a lot today, there are regulations of capturing people's data because that's what they're trying to do in London and understand who the person is, where they're going, are they afraid. We can look at them, we can read them. And we can see how they're walking, what car they're getting in, all these sensors and all these data points. It’s what I’ll tell your audience as well. AI is largely driven by data points, so think of data points, and AI is going to be using those data points to help them to change. Whether that's when you like to have your coffee for a shop, or when you like to get in the car, or when you need to drop your kids off at school, or what activities your school has, or all these different things they will be trying to feed it to understand: How do we give you the best experience?
Silvija: So smart cities I also think of as you have a problem where we now get so much data, that we are in some ways, well the danger I see in them, is that we may abdicate our local politics to AI. I think that every city in the world with any respect for itself is now wanting to be a smart city, but I think before we do that, we need to decide what does it mean for us? So, you know, what does it mean to have a smart city-London versus smart city-Oslo, or smart city-Bergen? Because these are very different places, different strengths, different local cultures. I think this is a really interesting area to look at the problem of hybrid AI humans, you know, so what is AI going to be doing for us, and what's going to be uniquely human? In the future. Do you want to talk a little bit about that? Where do humans fit in this applied AI loop?
Tom: Yeah. So, humans are always going to have a role in the loop. Or at least I think they will. I mean, there are talks of technologies like AGI, which some people say is 100 years away, other people say it’s 10 years away. And that allows cognitive thinking for a robot or a piece of AI software that will, you might have read about it, at Facebook, where they have started having its conversation, I don't know how true that is, or to what degree but when people look at AI.
Silvija: I just need to interrupt because I'm an AI researcher in my past life. I think that AGI is a completely, I mean, it's an interesting philosophical topic, but I think practically it's irrelevant.
Tom: Really?
Silvija: I believe that if we abdicate, if we leave it to the data, to decide on ethical questions and political questions for us, then we are giving up our purpose on Earth, you know. I think we need to want to do something with our lives, with our countries with our end. If we don't want to be the ones deciding, then, you know, what's the point of the whole thing? Then really, we are just cogs in this machine that's optimizing profits or efficiency for somebody, you know, a few, three, five people on the earth or something like that. So, I'm very, very keen on AI but as an applied tool and super keen on understanding that there has to be a human behind the steering wheel, wanting to do something with this tool. Unless we know what we want to do with the tool, the tool is useless and dangerous.
Tom: Yeah, well, I mean, it's going to be dangerous. And it's going to be very useful when people that are scared of AI, or think it's going to take their jobs, and it's not controversial, but we're upskilling all the time, there's more innovation. And if you're not ready to board that train to upskill yourself to use AI and understand it, you can't expect it to take your job because it's designed there for people that are high achievers, in my opinion. And it's going to be brought into a lot of businesses. But I always look at it like AI is there to aid you, to support you, to make you better, to give you better decisions, not to start taking things away from you. And I mean, there will always be people if there are 7 billion people in the world, I'm sure there are certain people out there that want to use AI for the wrong reasons, whether we like it or not, they're gonna have different viewpoints to what we have, whether we think AGI is right or wrong, or whether we think AI being used in industry is right or wrong, they're going to want to use it. Business owners and publicly traded companies are very big on profits, if they can find a way to get rid of people, they will, and that's what AI can unlock because it can unlock rapid learning 24/7-365 days a year. And it goes back to Elon Musk’s point, he said it's more dangerous than nuclear warheads. And he said, there's going to be more intelligence created by machines, almost 100 times to one than there will eventually be created by humans. Now, I don't know if that's true or not, because I see it as very derivative of what AI can do, and it can't create. It can't stand on its footing. And it doesn't pretend, it doesn't have bias. People say it has bias. But that bias is fed by humans. So, it's not their bias. It's our bias just collated.
Silvija: 350%.
Tom: Yeah, and I think it's an area. My key message would be, look, it's going to be used in everything and it's growing and everything. Look at any report, AI is getting bigger, it's being used. Last year in 2020, just in the USA alone, AI startups received $1 billion in funding each week, so that’s companies that are funding up to a maximum of around $20 million, a billion a week in $20 million companies or less. So you think of all those companies that need and some of them are only getting half a million. And that number that's just in the US, let alone the whole world, hotspots like Tel Aviv, Singapore, London, Paris, Berlin, Germany are getting very hot, especially in the manufacturing with all their car makers, and you think all these stats, that to me I remember reading that stat and I were like: “wow, this is exploding. This is not going anywhere.” And I think there'll always be room for humans. So just to finish off on that point. Technology doesn't drive change, people do. And that's what I've always been taught. And it's one of my, it's not my favorite quotes, but it's up there. Because we look at it from a media point of view where you send all these press releases, and people use the word digital transformation. And they do not talk around people, they do not say that they need people to drive these changes. If you do not have a person, if you do not have a data developer, if you don't have a data scientist, you don't have a solutions architect, you're going to do nothing for the day, sir, you're not going to create anything. And that's why companies get themselves into big ruts because they're trying to speak in plain English to your audience. And that's how I always tried to speak. If you've got a business that, say you're in retail, and they're having a massive logistics problem and may bring in a consultancy, whether it's McKinsey, PwC, or someone big like that, Gartner, to analyze this and to do all these changes. They might build you something which is called a testing center of excellence, or they might build in something where you can understand all the data essentially, that you can make use of it. And they might build you a framework or a system to analyze that data. But once those people go from that company, and say something changes in a retail shop, as a new product comes in, a new size, a new variant, a new color, a new pair of shoes, a trend because an influencer on Instagram decided to support a certain brand. You try feeding that to the algorithm or the data that this company's installed, and you don't have data engineers to update it and to edit it and tweak it. You're going to have a big problem because it's not going to understand and it's going to think I don't, and then it's going to get you poor results. So it's always, AI is powerful, providing you've got the right data, and I read somewhere, to give your audience another view, there's a thing called structured and unstructured data. So structured means it's very organized. So, you can read it. And you can see what it means, you can understand what categories it is. If you're looking at a house, it will be filtered by color, street, rooms, location to school, amount of cars that you can park on the driveway. All these filters, how many sitting rooms, whatever it is. If you have unstructured, it's saying: under the streets, you've got the number of bedrooms, so it's irrelevant, so it doesn't match up.
Silvija: It’s a text.
Tom: Yeah, and 95% or the 95-96% of data at the moment is all unstructured, which is bizarre, because we're creating God knows how much every day. And you can see where the problems lie. And you can see why people drive the changes, not just looking at AI. And I mean, it's a great marketing tool, because you can say, oh, we're AI-enabled - doesn't mean anything to me. Just because you're AI-enabled, I mean, I'm AI-enabled, I've got my smartphone, and it uses IoT, it uses all sorts of AI if I try and speak to it, hey Siri, or whatever. It's all data. So, it's very important. And that's the key message I want to tell people, people drive the change, not AI, driving the people. And that's why I say you shouldn't be scared because there's a digital skill shortage. People don't know how to use it. Sorry, I've gone on a bit of a rant, but it's very passionate to me.
Silvija: No, but I share it with you. So it's people with skills that drive the change and that's also my main worry about the future. Because if people don't understand this is just a tool. It's an incredibly powerful tool. But it is a tool. And unless we understand what the tool does, and how it can be used, and have some images in our heads on how it's being used, we are not going to be able to use it. And as you say, if you don't use it, then it'll take your job eventually.
Tom: If you’re not up it will take your job because you need to upskill, I mean, it's in everything. Just quickly. To give another example, I'm obsessed with the era of insurer tech and FinTech. So, financial technology companies and insurance technology.
Silvija: Insuretech as in insurance and fintech as in finance.
Tom: Yeah, so financial tech companies and insurance technological companies are changing it in a very way. So, if you look at it from a user perspective, nearly everyone I know needs insurance, and they need insurance for their house for their car, for the holidays, whatever it is, whether it's commercial or user. I always say a FinTech is going to move into holidays and it's going to move into insurance as well because they're collecting all that data. When you go shopping at the supermarket. What products do you buy? When do you buy it? When do you go to the store and buy it? At what time of the day? How many other items do you buy with it? Do you buy more regularly at a lower cost? Or do you do one big shop? Do you buy online from this retailer? And it's going to be collecting all this data from you as a user and from you as a purchaser. And understanding where you go, what you do, and tracking you through, if you've allowed it for your watch or your phone, and it makes it very easy for them to give you a better user experience. Because they can say, oh, here's a coupon for this shop that we know you like to shop at, because we've seen that you go there 20 times in the past six months or whatever. And here's back to our point earlier. Also, if you buy now, we'll give you a discount for flights to wherever it is you're going with three people that you always normally travel with. And you can see from an application view how it's going to affect the user because it's just creating this constant feedback loop of understanding them. But you don't get that feedback back loop without people that are very skilled, like the data developers to understand it. So that's what I yeah, another example.
Silvija: We have a cool example of AI being used in finance in Norway and Sweden. I don't know if you are familiar with Klarna?
Tom: Yeah, I am yeah, the payment system.
Silvija: They just are extremely good at connecting the data and making sure that they get a better payment.
Tom: Well, I'm not massive on Klarna so I'm not a big fan of what their business model is and who they are. But it's great how they’re using data, but they're essential, to me a credit company, and I don't like credit companies. So, I think it can be very intrusive on understanding the consumer, I think, but it's a very smart way of how they're doing it. They're building it in a very successful way.
Silvija: I agree, and you know, it is a part of every finance company, you know, that kind of risk, and depending on that risk. And then there is a Norwegian startup that I love called Aprila and they're a bank, and they provide small businesses with loan facilities, but they are much better at analyzing SMB. So normally, banks are quite cool with small businesses, it's high risk and relatively low gains compared to big businesses. And Aprila has made it into their specialty to understand the quality of the business from random open data you can find about those businesses and thereby being able to provide loans to these companies at a competitive rate, which I think is a wonderful business model. Listen, Tom, so we talked about AI, in agriculture, in climate, in engineering, in production, in travel, and in transportation, as in airlines, airplanes, or cars, as in digital twins, we talked about smart cities. We haven't touched smart homes. We'll save that. We have talked about AI in finance and insurance. But we haven't talked about AI and education and health. I think we have enough to talk about for at least another half an hour. But we have five more minutes left. So I'm going to give you a choice. Do we want to say something about smart homes, education, and health? Or do we want to give people some advice on where to go next? What should they read, in your opinion? Or do you want to do both?
Tom: I think I'd like to focus on healthcare because it's such a big point. But where to go next as well. I mean, education is a huge topic. And I'd suggest people go and download the reports we've got on AI Journal and the work you're doing as well being an education company. And VR is going to be huge for education. I'll leave that point there because it's immersive, you can understand it, kids can be in a room, and they can see what's going to happen next. And they can understand how things are changing in a very practical onsite view. But that’s also very crossing to healthcare with such a big area. And all of these intersect. That's the great thing with AI right because if you're using IoT smart homes, that data can be used for your healthcare provider to understand your eating habits, how healthy you are, your body when you go to sleep, are getting enough sleep, all these associated with your health and whether you're at higher risk of cardiac diseases, or whether you aren't getting the proper sleep and all these feedback loops that can be created. And I mean, VR, we've got an example in the report of it being used in surgery. You can put a headset on, and you can see through a body, and obviously, I don't know how patented it is and how available to general use in hospitals it is. But it allows you to scan the whole body and give you an AR layout of that person to detect where the problem is. So you can put a sensor in them, and it will highlight where you can see and what's going on with that certain organ. And I think it's very important in healthcare to understand, it’s probably where it's going to be the biggest game-changer because it's going to keep people alive, you can use AI to detect where cancer is going to happen. So they can run it through loads of different simulations. And they can use all those data points on you as a person to understand is this going to happen, or what to look out for. So when a doctor is doing an assessment, where should I be looking? Office are a bit of a slow mover, but where should that person be touching, feeling? What kind of questions should I be asking to get to a diagnosis, not just for cancer but with God knows how many other diseases, Parkinson's, or all these other unfortunate things that happen. And with pharmaceuticals as well to discover new drugs. So with COVID, AI played a massive role in understanding how the vaccine, the COVID vaccine, could be manufactured and what the right geometrics combination was to create a vaccine that can save millions of lives, proving what you see on the news. That's all we're using AI to understand, where should I be? Where should this be going and what data should we be using? And that's a key message I'd always tell people, AI does not need to be complicated. It's a prediction tool. It's an accuracy tool. It's something that is a capability, I wouldn't even class it as a technology. It's a capability that if you got a problem, it could predict it for you at a very high level or a very low level, using something or one of the buzzwords that you hear “machine learning,” which is essentially just learning datasets as quickly as possible. So, to learn more about it, and that's what they do in healthcare, VR, and everything else with smart devices, I would honestly check out the AI journal weekly, we got lots of free resources. Check out what Lørn.Tech are doing. And there's a lot of cool areas, especially if you want to get more technical and understanding it at a very deep level, places like IEEE that you can go to for journals. And looking at it from an influencer side, a big area that I use is LinkedIn. I honestly get more insight from LinkedIn than I do from Forbes or TechCrunch, or Wired or any of these other kinds of science journals. To look at who these people are, you can follow people, whether it's people like Danilo McGarry, or Renes Rakishev who's great on data, or someone like Kieran Gilmurray, all these people, and we've got a free resource we're building into the website, that's where I'd go to check out more or use Twitter because they talk about it in much, much more depth that we don't have enough time for today. But also, application way, but also considerations like which we didn't get to touch on today. But the ethics, which is a huge area, and bias and all these different areas, because ultimately, at the end of the day, I'm someone that sees it from a top-level, they're someone that's working in very big corporations and deploying it and using it for a lot of clients and building those models. So that's where I'd recommend going. I'd also recommend stopping telling yourself it's a tricky topic. It’s is a very simple topic, if you need it to be. It’s a prediction tool. It can be used in anything and any way. The next time you're thinking about how it's being used, look around you look at your phone, Siri, look at when you're going to the next shop, look at when you're getting in your car, all these parts somewhere either in its production or in its current state, we'll be using AI and we'll be there to give you ultimately a better experience. It's not there to take your jobs, it’s there to enhance your job, it’s there to give you a better experience, and it's there to help you and I'll always stick to that message.
Silvija: Tom Allen, I think I'll leave that to be the famous last words. Thank you so very much for coming here to Lørn with us and inspiring us.
Tom: Really appreciate it, thank you.
Du har nå lyttet til en podkast fra Lørn.Tech. En læringsdugnad om teknologi og samfunn. Nå kan du også få et læringssertifikat for å lytte til denne podkasten på vårt onlineuniversitet Lørn.University.
LØRN AS, c/o MESH,
Tordenskioldsgate 2
0160 Oslo, Norway
Bibliotek
Om LØRN
© 2024 LØRN AS