Why driverless taxis still need a human (for now)

0:02Sky News. The full story first.

0:11The UK's first autonomous taxis have arrived, but they still need a human sitting behind the wheel. And this is why.

0:24Who says Halloween only lasts one day? With Wayfair, one night of Halloween becomes a whole season at home. From larger-than-life yard decor and moody lighting to festive accents for every room in the house, Wayfair has what you need to make Halloween feel like magic all in one place. Because the best Halloween memories aren't made in just one night. They're made all season long. Shop Halloween decor now at Wayfair.com. That's W-A-Y-F-A-I-R dot com. Wayfair, every style, every home.

0:59Hi everyone, Neil here and whilst I very much consider myself an early adopter when it comes to tech, there are a few things I don't want to hear during a journey in a driverless car. And this is fairly high on the list. I just watched an autonomous car make a mistake on London's roads. And the safety driver had to take control. But the plan is that eventually there won't be one there. Still, making sure everything works as you'd hope is what trials are for. But that human safety driver is also what makes this interesting. The AI does the driving, the person supervises it and corrects its mistakes.

1:37Until, that is, the machine is considered good enough to manage alone. But this is a very small trial. 15 cars, each with a human ready to intervene. Which raises the obvious question. Are driverless taxis really arriving in London? or is this mainly a chance for the companies behind them, Wave and Uber, to show off the technology? How close are we to machines driving by themselves? And what does this tell us about the future of work? Roland Manthorpe, this guy's technology correspondent. Roland, just explain exactly what has just been launched here in London. So for the first time ever, it's possible to hail an autonomous vehicle through the Uber app.

2:18The vehicles are made by a British company, Wave, and they've partnered with Uber to offer this service. But it's a bit of a funny one, right? Because they're driverless cars, but there's a driver. Sorry? The driver is, yes, I know, I know. There's a safety driver, and they sit there with their hands just beneath the steering wheel in case anything goes wrong. So, yeah, they're autonomous vehicles. But I'd say this is probably more in the nature of a trial and a learning experiment than the real thing hitting the road. OK, so you have been in the back of one of these sort of driverless cars. Tell me about it.

2:51How was it? I mean, it was a crazy, crazy, crazy journey. Everything that could have happened, happened. We drove down Whitehall to Parliament Square and there were people all over the roads. There were horses going the wrong way down the road. There was a car with blue lights going really fast the wrong way down the road. and the wave driver handled it all pretty smoothly until we came up to parliament square and later we worked out that what had happened is that just stop oil had reappeared for the first time in quite a long time and started gluing themselves to the street outside parliament and the police sort of almost at that moment blocked off the road and the wave got confused and instead of following all the traffic which was now being diverted it wanted to carry on its route and it started to head ever so slowly towards the blockage while cars were coming around the outside.

3:46And I've looked back at the footage which we got from the inside. You can see the steering wheel is turning in the wrong direction as cars come around the outside and the safety driver had to intervene. There was this kind of beep, beep, beep, beep, beep and the safety driver had to intervene. And it was a really, really, really surprising moment. Surprising or scary? It wasn't really scary at the time because we were at such a slow speed ultimately. And I'm not totally sure what happened in the car and Wave don't really discuss these incidents. They take them back and learn from them, but they won't tell you exactly what happened.

4:20It was very unlikely that something serious would have happened because I think the wave would have jammed on the brake. In any case, the safety driver was there. It was afterwards that I thought, wow, that could have been dangerous. I was shocked because I've driven in these things. I think they're brilliant. I think this is going to be a really important technology. I've been in a wave before and everything went really well. You very quickly get used to it handling things quite comfortably. And there's just something about seeing that autonomous steering wheel turning in the wrong direction that, yeah, it spooks you.

4:53It does. And I guess it can shake my faith in the whole project a little bit. It does make you wonder, can driverless cars really handle London's roads? Without meaning to sound too simplistic about it, isn't that an example? isn't that evidence that quite clearly they don't? And actually, this should be something that we should be a little bit more concerned about when it comes to, at the very least, the sprawling nature of the roads here in London. They're clearly not perfect, but neither are humans. I said everything happened on this journey. We actually saw a crash right in front of us. A taxi driver tried to nip into a space that just wasn't there, and the bus hit his car ever so lightly, and he was jumping out and shouting and getting really cross.

5:33And the wave, by the way, handled all that very smoothly, navigating this totally mad incident. I suppose up until now, if I'm honest, I've always thought that the idea that driverless cars couldn't handle London's roads was really sort of UK chauvinism. Everyone thinks their roads are difficult, but ultimately they're just roads. And the data that we've got coming out from Waymo, the American driverless car company owned by Google, is incredibly impressive. I mean, they are now operating in multiple cities in the US. I think in total, they've driven something like 220 million miles and having 80 % fewer injury crashes than human drivers.

6:12That's compared to America's districts and they crash a lot more than us. But even so, you know, we're seeing a lot of data come back that this thing really works. But seeing yesterday, it was a reminder, London's roads, there's just so many unpredictable things can happen. And yeah, it made me wonder. It made me wonder. A lot of this technology is kind of already out there. I mean, I think plenty of us already know someone who has a button in their car that could probably do most of what's going on within these robo-taxis. How do they work? How on earth can a car judge how to drive correctly, properly and safely?

6:46There are lots of answers to this and people are pursuing different approaches. But just as a very sort of general principle, the whole way that AI and all this stuff works is that it just sucks up an absolutely gigantic amount of data and then learns from it. And so we have a lot of, lot of, lot of video and driving miles and pictures telling us what roads are like. You give the machine that data and it can learn from it. That approach on its own hasn't proved sufficient yet. So you also need to give the machine the equivalent of human senses. Actually, in some cars, you give them superhuman senses.

7:30You know, you give them cameras to mimic our eyes and you give them microphones to mimic our ears, but you also give them LiDAR and radar and other senses. And when you see what they can see, you realize like, wow, they've got much, much, much better senses than us. So how do you see what this car is seeing? So LiDAR is a good example. It's a radar system. They used to be quite expensive, but now they're made quite cheaply in China. you can look at a lidar point cloud the output of it and you quickly gather that it's seeing much more clearly and in much more higher fidelity than we do because essentially what you're doing is you're giving the car senses through the sensors and you're giving it knowledge through all the data and it does work really really really well it's just this thing right you can't get it right most of the time in a car.

8:24You've really got to get it right almost all of the time or all the time to be safe. Are we ever going to see driverless cars? Cars without a driver? No one sitting in the front seat, their hands in their laps, ready to grab the wheel if the thing gets it wrong. Is that the direction of travel? They're already there. Multiple, multiple US cities, San Francisco, Miami, Austin already have these cars. I feel like there's no way this technology doesn't work in some form or other. I think to me, the question is scale. How big can they get? How many cars can they replace? And this comes also to the economics of it all.

9:02Outfitting a car with sensors all over it and it's specially made and has special training on it and everything like that is super expensive. How many cars can you get on the road if you're going to have to do it that way? One of the interesting things about Wave, the British company that is doing this partnership with Uber, they're trying to create something which is a bit more like us in that it can see an unfamiliar road and drive it because it has a machine-like brain that can see situations and understand them in real time. And yeah, you know, I think they've been making tremendous advances.

9:38It's a really, really interesting and impressive company valued at many, many billions. But yeah, as my experience yesterday showed, They're not quite there yet. They're driverless cars, but they still need a driver. What then is the point of this technology? What is the problem that is being solved? I'm not sure that I care very much about how I get from point A to point B as long as I get there safely. And there are plenty of ways in which I can do that right now. I think there are two really big answers. One is that there are a lot of people injured and killed on the roads every year. So far, the data we have, which admittedly comes mostly from the US, which is a more dangerous place to drive, shows that these cars are a lot safer.

10:21You might say in some sense that we actually have a moral obligation to introduce them, that this is just better technology that can really save lives. So that's one thing. And the other thing is, is that honestly, I know this might annoy some people, but I is that if you can really think about it, driving is quite a massive, well, actually a colossal waste of time. All those people, especially in London, right, just sat there in traffic with their hands at the wheel, unable to do anything else, just stuck there. I think that we look back at jobs that were very important in previous times and we think, wow, can you believe they did that?

11:01What an unbelievable waste of time for everyone involved. You know, a good example is lift operators. When lifts were first introduced, they were manually operated for quite a long time. And when automatic lifts came in, people were really scared. And the lift operators were furious, there were big strikes. But I think it's that now, within the fullness of time, we can look back and say, having someone just stand in a lift to press some buttons does not seem like the best use of human time, really, does it? And I take the point. I suppose, though, the 100 ,000 licensed taxi and private hire drivers in London alone would suggest that their job is somewhat more complicated than sitting in a lift and taking people up and down.

11:39I mean, we know that Wave have partnered with Uber for this particular trial. You're not telling me that Uber wouldn't absolutely adore to be able to get rid of a particularly sizable chunk of their expenditure if they could have a taxi fleet without drivers, more money in their pocket, more profits for shareholders. Oh yeah, and of course, like it's very irritating to an Uber driver to hear me take this like Olympian grand historical perspective. I mean, yeah, this could be a very major disruption that could be very painful for a lot of people. I think one thing that I would say about it though, is that driverless cars seem to be coming much more slowly than other kinds of automation.

12:19That's due to lots of reasons, both to the technology, but also the way the companies are rolling it out. So I think the city that has the most driverless cars at the moment in the world is San Francisco, and it's got about a thousand. And the number of cars that Wave are introducing with this partnership with Uber is in the tens. So this is not arriving at an enormous, enormous, enormous speed like the way that software AI is. So if things keep going at this pace, we are probably going to have quite a lot of time to adapt. Of course, that won't make it any less painful to people affected, but it might make it less shocking than some other kinds of automation that are coming through software AI.

13:00There does, though, seem to be something of a pattern emerging when it comes to those kind of industries, those jobs, those roles that are proving to be most vulnerable to artificial intelligence. We've talked about drivers, private hire, or taxis. We're also seeing drones being brought in to replace delivery drivers. You know, call centres, left, right, and centre, are closing in favour of AI. Now, these are all pretty low paid and viewed as low skilled jobs. I mean, we're talking about the loss of tens, hundreds, perhaps even millions of jobs that are essentially entry level jobs that most people could do.

13:34And it doesn't seem to be affecting the better off as much as it is affecting the worse off. Oh, honestly, I think you're totally underselling it. And I think it is affecting the better off massively. I'm sure they'll find a way to get around it. but I would not feel very comfortable if I was a lawyer right now or an accountant or really any professional services job looking at what AI can do. I think that everyone must feel that kind of slightly concern in the pit of their stomach, right, when you see how good these systems are. But I suppose, like, can I just be annoying again and take a sort of slightly broader perspective here?

14:08What is the reason why there are so many what you call low-skilled drivers now in London? Why are there so many Uber drivers? That didn't used to be the case. What has made that possible is automation. And what was automated was navigation. It used to be black cab drivers really had the monopoly on driving in London, and they had the knowledge. And you know, without the knowledge, it was really hard to get around and navigate people efficiently. Suddenly, you get automated mapping, and anyone who can drive can get in the car and roughly work their way around London. That has been an enormous change.

14:43and if you speak to black car drivers, they're not very happy with it. It's massively driven their wages down. So yeah, I think in a sense that change is a constant here and we're seeing recurrent waves of automation. And of course, it's incumbent on society to deal with that in some way. I mean, this is not very reassuring, but I think that there are other more aggressive types of automation that are coming way faster and that we should probably be worried about even more than this. Roland, as always, many thanks. Peddito. And that is your lot for today. Do get in touch with your thoughts. Many of you, I imagine, are drivers.

15:17Would you be happy sitting in a car without that human safety driver? Do let me know. Why at sky.uk is the email. We're back tomorrow.

15:31Who says Halloween only lasts one day? With Wayfair, one night of Halloween becomes a whole season at home. From larger-than-life yard decor and moody lighting, to festive accents for every room in the house. Wayfair has what you need to make Halloween feel like magic all in one place. Because the best Halloween memories aren't made in just one night. They're made all season long. Shop Halloween decor now at Wayfair.com. That's W-A-Y-F-A-I-R dot com. Wayfair, every style, every home.