AIs on the road

0:00Welcome to Tech Life on the BBC World Service, the programme about technology and how it affects our lives wherever you live in the world. I'm Chris Vallance. This week, how would you feel if an AI in your car watched you drive, raising the alarm if you fiddled with your phone or got too close to the vehicle ahead? How would you feel if it told your boss how safely you drove? And sticking with AI, we'll hear what children think of the artificial intelligence-filled future they're growing up into. Children are the group who are likely to be most impacted by advances and AI technologies, but at the same time, they're the group who are least represented in decision-making.

0:39And getting medical supplies into hard-to-reach places in Central Africa with a drone that's cut its teeth delivering the mail to Scottish islands. For the first time, you can deliver long-distance, relatively high payload and do it completely autonomously.

1:18OK, well, let's get this show on the road. Imagine you're driving a van doing a steady 50 miles an hour when you decide to take your eyes off the road and fiddle with your phone. How would you feel if your vehicle suddenly instructed you to do this? Please put phone down. Or what if your van told you to do this? Please keep eyes on the road. Well, those warnings came from a system currently fitted to thousands of delivery vans and fleet vehicles. Made by a company called Netrodine, the system uses cameras and an onboard AI to assess how safely the vehicle is being driven, warning the driver of hazards and letting managers know about the safety of their drivers.

2:03The techs in active use in North America, Australia, New Zealand, India and parts of Europe. It's designed to enhance safety, but could it also feel a bit intrusive? I've been speaking to Durga Naidenghadi, Senior Vice President at Netrodine. Our mission at Netrodine is to save lives, right? It's actually to use technology which is available today to make roads safer, to make vehicles safer, to make drivers safer. What we do is to use technologies like artificial intelligence, machine learning, etc., all of these buzzwords that are around today, to basically detect using camera signals, how your driver and your vehicle are actually doing on the road.

2:56So let me give you an example. We have a device which has a camera which looks outside of your vehicle and basically monitors everything that your vehicle is doing on the road with respect to everything outside, right? We also have a camera which looks inward into the cabin which looks at how the driver is actually behaving. So the outward camera has the ability to detect speed signs, stop signs, traffic lights. How closely is your vehicle following the vehicle in front of you? Is there a pedestrian whom you're likely to hit, right, coming in front of you? All of these are detected by AI, processed on the device itself in real time.

3:43and whenever there is such a possibility, either you've exceeded the speed limit or you've jumped a stoplight or, you know, there is a potential collision, in real time there is an alert which is issued in the cabin so that the driver can take, you know, corrective action in real time. So this is a little like having a driving instructor next to you, if you like. Absolutely, a driver coaching you basically, right? And I'm glad you asked me that, Chris, because a lot of times cameras get equated to surveillance and surveillance is not what we do. The only time we actually use the camera to upload videos is whenever there has been an infraction.

4:25For example, if you've jumped a stoplight or if the vehicle sees you talking on the phone, it raises an alert and a minute long video of that alert gets transmitted through the cloud to potentially a safety manager's portal or a fleet manager's portal. You know, someone who has the authority to oversee the behavior of a driver in a large fleet. So you've got a camera that's pointing outwards. You've also got a camera looking inwards into the cab. And that there is looking out for behaviors that might be dangerous. So you mentioned looking at a phone, perhaps anything else? Absolutely. So we look, firstly, we detect seatbelts, right?

5:07Is the driver wearing a seatbelt? Not fundamental, right? Then we look at whether the driver is distracted. Is he or she actually, you know, moving head many, many times? Drowsy, you know, we detect yawns. We detect fatigue. If you've driven long distances, you would realize, right, you either start blinking too fast or your blink lasts for much longer than it should. obviously drowsy and fatigue are the most serious ones for us so though you know we categorize those as really severe alerts so that people can take correction action immediately so through these in-cab real-time alerts we help the driver avert a possible incident not only do we do that now Now, assume that, you know, you're managing a fleet of, say, 100 vehicles.

6:01You probably have 150 drivers driving them at different points of time. Using the video that we capture of this driving behavior by drivers, we actually give drivers a score, what we call a green zone score. Green zone is all about positivity, right? How well are you driving? So not only do we track all these incidences of where the driver could have driven better, we also track whether the driver's been driving well. For example, if you've driven eight hours without any infraction, you've done a good job. It doesn't mean that the technology should always use the stick on you, right? The carrot works very, very well.

6:43So you're talking about its use in fleet contexts. Give me examples of the kind of organizations that are using this technology and how widely deployed is it? So we now have three and a half thousand customers worldwide. And our customers range from some of the largest fleets carrying heavy, heavy trucking fleets, if you will, right? Long haul cargo carrying, oil and gas carrying, hazardous material carrying fleets, right Right down to the friendly neighborhood man who delivers groceries to your doorstep. The largest e-commerce company in the world is our global customer. That would be Amazon.

7:29I know you probably can't confirm that. I don't confirm or deny. I was having a look on social media, looking at some comments from people who said they were employees of a very large e-commerce company that used your tech. And there were mixed reviews, I guess I'd say. Some of the drivers felt that it was good. They could be confident that if there was an accident or incident, the evidence was there, that their managers would see a clip of what happened and they'd be exonerated. Others were very worried that they were effectively being spied on in the workplace, and they were worried about things like, is it listening to my conversations?

8:12What would you say to those who have concerns? It's an absolutely natural reaction. And we face that every time we deploy our technology into a new fleet and a new driver starts driving. the first human reaction is I don't want this thing looking at me firstly and then making noises when I think I'm a great driver. So the first reaction from drivers is always resistance. But we find that very quickly, you know, they understand as soon as you cross that threshold where they know that the intent is not surveillance, but it is, you know, assistance, then their whole attitude towards the technology changes.

8:58In fact, I've seen drivers go through phases of being, you know, absolute opponents to being proponents. They actually do word of mouth nowadays and say, guys, get your fleet manager to look at this. And very briefly, you're the boss, or at least in the very upper echelons of the company. So nobody to report to in the sense of your driving. But do you have it in your own car, just for your own safety? Yes, I do. Yes, I do. In India, I do. But as you can imagine, driving in Bangalore with a device like this, unless it is tuned for the Indian roads, which it is, can be quite eye-opening, let me put it that way.

9:41But I do. I do. And that was Durga Nadengadi from Netrodyne. And if your workplace uses their tech or something similar, or you just have an opinion on what Durga said, do get in touch. Our contact details are coming up in a moment.

10:07This is Tech Life on the BBC World Service with me, Chris Valance. and, of course, with you, our listeners. Alvin Banderas, who tunes in or downloads or streams from Manchester in the United Kingdom, emailed us after hearing us talk about vintage technology. Alvin says he has a huge collection of pocket and handheld electronic organisers and computers from the 1980s and 1990s. An example, Alvin says, is the Sharp PC 1500 pocket computer. It became popular from 1982 with its optional miniature printer and plotter for the businessman and engineer. Well, thanks, Alvin, for that trip down memory lane to the era of big hair and bigger shoulder pads.

10:54Well, from emails about vintage tech to tech that's absolutely cutting edge, we've been asking you to send us your questions about quantum computing. Is there something you'd like to know about this fascinating but highly complex technology? Send us a question and we'll put a selection of them to quantum experts on a future edition. Well, whether it's about old tech or the computers of the future, you can contact us by email. Techlife at bbc.co.uk is our email address or send us a WhatsApp or voice note. The number is plus 44 330 1230 320. Do remember to include your name and where you live. And still to come, what are children telling adults about AI?

11:51And now, tech life is taking to the skies.

12:02Well, that's the sound of the Mark II Ultra drone from British manufacturer Windraces, and it will be heard all over parts of Central Africa in the not-too-distant future. The company has teamed up with Aviation Sans Frontier to use the heavy lift drone to deliver vaccines and medicines to hard-to-reach places in the Democratic Republic of Congo. David Holden is Africa operations lead at Wind Races and he's about to travel to Malawi, where the company is also working on building a drone hub. He told me more about what his company's drones will be doing. So we're going to be delivering humanitarian aid.

12:40So we are looking forward to going to Democratic Republic of Congo in October and we'll be flying our first missions in the province of Mai Ndombe, working with Aviation Sans Frontieres to deliver humanitarian aid to some of the remoter parts of that area. We've heard a little bit of the sound of one of your drones, David, the Windracer's Ultra Mark II drone, I think. One of my colleagues said it sounded a little bit sort of, you know, like a kind of World War II plane. And in a way, it sort of looks a bit old-fashioned. It's got two propellers. Tell me about your drone. So the drone is not something that you'll be used to flying in the back garden on a Sunday afternoon.

13:25This is an aeroplane. it has close to a 10 meter wingspan it's roughly six meters long if you see it in the sky it would look not dissimilar to a normal general aviation aircraft it's a big it's a big drone the drone is flown completely autonomously and i think the best way of thinking about that is that you're not flying it with a radio control or with sticks like you may be used to the drones that you fly at home, it is essentially sending a computer program to the aircraft, and the aircraft then will carry out the mission, the profile that you've sent it. And if you want to change that mission, you can at any time.

14:08But the aircraft will continue to fly in detail the mission that you've already sent it. So when you see it, you see what looks like an airplane, you just suddenly realize that there's no one sitting in the front. Of course, that's a very big difference and represents a very significant advance in technology over the propeller planes of the past. Tell me about what it can carry. Yeah, so it's a large aeroplane. So you've got a payload of 700 litres and a payload weight of up to 150 kilograms. So you've got quite a large space there for which you can carry really a range of items. It can carry quite a long way.

14:51So we've got a thousand kilometre radius on the aircraft, which means for the first time you can deliver long distance, relatively high payload and do it, as I said, completely autonomously. Describe a typical mission for me when you start flying with Aviation Sun Frontier. Yeah, so what we expect to do with Aviation Sans Frontières is to identify certain locations, very remote locations, which will require humanitarian aid to be delivered. Now, obviously, they won't have a runway for the aircraft, so the aircraft will deliver those through a dropping mechanism. And the way the aircraft works is it flies extremely accurately to a predefined waypoint at low level.

15:39And then the door is opened and the boxes are released. And the boxes can be delivered either by parachute or they can be delivered without a parachute. But in the old days, you probably have an image of sort of people chucking crates out the back of a Hercules. And that's not the way it works here. These can be delivered at low level, at relatively low speed, and really with extreme accuracy. Why not use a truck? What's the advantage of using a drone? Or indeed, why not use a piloted aircraft? In terms of a truck, there'll be many of your listeners who understand the geography of DRC much better than I do.

16:16And it can be an extremely difficult place to get around. So actually, sometimes truck can take days, if at all, especially in the rainy season. And often the alternative is a mix of river and a mix of truck, but very difficult and very time consuming. Obviously, we can be much quicker because we're just direct point to point. Also, in terms of why not crewed aviation. So crewed aviation is far more expensive and you're having to pay for much larger, more expensive aircraft, crude aircraft. and normally aircraft that are not well designed to drop accurately the types of cargo that we're planning to drop.

17:05And separately, you're also building a hub for drone operations in Malawi. What will the purpose of that hub be? What will it let you do? And what does a drone hub look like? So, yes, it's really there to be a facility where we can demonstrate the aircraft. There really is across Africa a large amount of interest in the proposition, in the aircraft, in the capability. But I think like most people, people want to see it actually operationally demonstrated in the African environment. So we decided to set up and we built a bespoke hangar at Kuzungu Airstrip. And we're going to base ourselves there.

17:49We're going to get a high degree of experience ourselves in the African environment, but also it gives us an opportunity to invite a number of potentially interested parties to come and see the aircraft and also actually design missions that they want to see operated in their country. I mean, you've spoken about operating around the world. I mean, you know, here in Scotland, your drones have been helping deliver the mail to some of the islands, Orkney and Shetland, where there's a big sea journey otherwise. Yeah, that's correct. So we've operated a large number of flights in Scotland, and particularly in Orkney, particularly in Shetland.

18:31But we've also flown missions in Antarctica, and a lot of interest in using the aircraft for what we might call advanced environmental surveys in that type of environment. And the aircraft are headed for both Alaska and Norway. And we've also flown a lot in Eastern Europe, where a lot of emergency resupply and intelligence missions have been flown. That was David Holden of Windraces.

19:06Now let's return to the subject of artificial intelligence. Across the world there are fierce debates about what rules should be in place to control this powerful tech. But are children's voices missing from these conversations? After all, they will be growing up in a world where intelligent machines are the norm, not science fiction. The Alan Turing Institute is the UK's national institute for data science and AI. Its experts recently spoke to children, their parents and teachers as part of research into the tech's impact on the young. TechLife's Graham Fraser met up with Dr. Barry Aitken, a senior ethics fellow at the Institute in London, and she told Graham what they found out about AI's role in children's lives.

19:52Children of all ages are now interacting with AI on a daily basis and that's everything from infants and very young children playing with smart toys that adapt and personalise through interaction through to older children viewing content online where AI prioritises and filters content for particular users. Then of course there's AI on social media which is shaping the information that children see and interact with as well as mediated relationships and friendships. But there's also lots of ways that AI is being used to inform decisions about children's lives, such as in education or healthcare or housing.

20:25AI is really a part of the way decisions are made across lots of different areas that impact children's lives. In our team and the children in the AI team at the Alan Turing Institute, we believe that children are the group who are likely to be most impacted by advances in AI technologies. But at the same time, they're the group who are least represented in decision-making around the ways that those technologies are designed, developed and used. but also in policymaking and regulation. So it's really, really important that we pay attention to the particular needs of children and young people and also the particular risks that AI might present for children and young people.

20:56Now, I've heard you say before that AI is affecting us all, but the one community of people who it will affect the most are the one community that we aren't hearing from, and that is children. Absolutely, yeah. And I think it's really, really important that children's voices are heard in discussions around the future of AI. Children have particular needs. they also have particular rights and those rights need to be upheld in the ways that technologies are designed developed and used and also in policy and regulation but all too often you know adults working in this space even even adults who are you know whose aim is to you know develop safeguards or develop policies that might protect children from the risks we often start from assumptions of how children might use those technologies or interact with those technologies and those assumptions are often based on adults experiences or the ways that adults use those technologies but we know that children are often early adopters of technologies they also often use technologies in ways that are very very different from how adults use those technologies or even how the designers and developers of those technologies anticipate or envisage that those technologies will be used so it's really important that we engage directly with children and young people to understand what their actual experiences are as well as what their actual needs are and what the actual impact of those technologies are in children's lives.

22:10Now I'm 40 years old and I'm of the generation where we grew up with computers and it became second nature to us but my parents for instance they found a little harder and it took them a bit more time. Children who are now when they grow up what their experience is going to be like because of AI that's going to be so ingrained in everything they've done in their lives? Yeah and this is where we're seeing particularly with advances in generative AI, it's changing a lot of ways that we do things. So children now are more likely to search for information on something like ChatGPT or generative AI model rather than a traditional search engine in a more kind of conversational way.

22:51There are particular risks there. These tools are really not, you can't rely on them for accurate, trustworthy information. They're not really designed for that purpose. Also, there's potential impacts on critical thinking in that when we, if you search for information through a generative AI model, it kind of does the process of filtering that information and presenting that to you. Obviously, search engines are in no way perfect, but when you search on a traditional search engine, you'll get a whole list of possible results. You can look through, you can look at the sources, and you make your own decision about which information to select and use.

23:22So there are real concerns on what the impact of that might be. There's also, increasingly, we're seeing generative AI being used in the form of AI companions or AI friends, and these are increasingly marketed at children and young people and these are AI personas, essentially chatbots which can be very, very personalised. Users can design their AI companion in relation to the physical appearance they want but also the character traits and personality traits that they want this AI companion to have and these can be very deceptive, they can be very addictive and we know that children and young people are increasingly interacting with these AI companions but as yet there really isn't enough research to explore what the impact of that might be on children's emotional development, on their social development, and on their future relationships.

24:08So I think we really urgently need more research to engage in these topics and look at it from children's perspectives around what their actual experiences and concerns are. What would you say are the... Because I know at the Alan Turing Institute you study this a lot and you've looked into the experiences of children and AI and you have done for years. what would you say are the big hopes that children have with this technology? Yeah, I mean in our research we find that children are very excited to engage with this topic and they often come up with very creative ideas around how AI can be used well in the future.

24:41Things like they want to see AI used in education to personalise education or to open up new opportunities in education. We find consistently that children are really excited about the ways that AI could be used to support children of additional learning needs or neurodivergent children and to adapt learning materials and learning approaches to suit the individual needs of different children and different learning needs. We also see lots of excitement around AI in research, particularly medical research or in relation to healthcare. And children are also really excited about the opportunities to use AI in relation to the environment and sustainability.

25:15So quite consistently we find children excited about the opportunities to use AI to model climate change or to use AI to track and monitor endangered species. That's Dr. Barry Aitken from the Alan Turing Institute, talking to Tech Life's Graham Fraser. And I wonder what our younger listeners made of that. Do get in touch.

25:39Well, we're almost at the end of the show. And just a reminder that you can still send us your thoughts on anything you've heard on today's programme or tell us about something you'd like to hear in the future. We promise a real human being will read your email. Do get in touch. You can WhatsApp us on plus 44 330 1230 320 or email techlife at bbc.co.uk. Please tell us your name and where in the world you're writing from. Today's Tech Life was produced by Tom Quinn, edited by Monica Soriano and presented by me, Chris Vallis.

26:24Two-Horns Subtitles