Is London ready for driverless taxis?

0:00This is The Guardian.

0:09Today, driverless taxis are on their way to London. Should we be afraid?

0:22So it said, look, you match with the Waymo one minute's... Confirm ride? Yes. I don't know if you can tell, but this is me, beside myself with excitement. I'm standing on 6th Street in downtown Austin. I'm full of tacos and margaritas and one of my dreams is about to come true. Hello driverless car. It says good afternoon Helen.

0:51When I went on holiday to Texas recently there was really only one thing on my bucket list. Sorry Lyndon B Johnson Presidential Library, apologies Blanton Museum of Art. All I really wanted to do was get in a driverless taxi. And it didn't disappoint. I can't believe it! So now it says finally relax and thank you for riding on us. Takes corners quite quickly. Proceeding on Abasoda Street. There's nobody driving! Waymo driverless taxis have become a common sight on the streets of US cities in recent years, ordered via the Uber app. And now, they're coming to London. This is one of the things that I thought would be the future when this happened.

1:43But how will these ghost drivers cope with our medieval streets, our soggy weather, and our total absence of anything resembling a grid system? And the biggie. Can we trust them?

2:01from the guardian i'm helen pitt today in focus robo taxi

2:12johanna bouya you are a senior tech writer here at the guardian and you live in new york now but until pretty recently i think you were a resident of san francisco and so you're an old hand when it comes to these driverless taxis. Do you remember what it was like when you first got in one? Yeah, back in 2015 or 2016, when I first started writing about self-driving cars, and there were 30, 35, 50 self-driving startups out there. Google has said it wants its cars in the hands of consumers by 2017. At Volvo Cars, we have to object them. Mercedes-Benz has been the leading car company when it comes to autonomous driving.

2:53It was major automakers, tech companies, random startups that just really had no business building self-driving cars, testing out their vehicles, which sounds wild now because there are maybe four that people know of. And so I can't remember exactly what automaker it was, but it was very much like a test car. It didn't look like a polished vehicle. It was very clunky. There were things all over the vehicle, but I think it was in Singapore. I was testing a vehicle on the streets of Singapore and went around some Singaporean commercial park. And it was a little bit terrifying because I think if anyone who's listening has been in a self-driving car today, the ride should be pretty smooth.

3:38Back in the day, those rides were not smooth. It was very jerky, a little bit dicey when you're in the car, but it was still really exciting. It felt like there was a ghost behind the wheel. It looked like something out of a weird sci-fi film. Nothing like what it feels like and looks like today. Yeah. And you can tell a mile off that Waymos are not ordinary cars. And Waymos, of course, are the ones who are going to be heading to London. They're white. They're electric powered. They're Jaguars. And they have like this big spinning camera radar thing whirring away on the roof. And then there are sensors all over the car.

4:15But how does this technology actually work? because I don't understand it at all. How on earth do these cars drive on their own? I mean, at the risk of oversimplifying it, and I always feel bad when I do this for self-driving experts who are listening to me, but for everybody else. Yeah, assume we're all idiots. Yeah. The tech stack is basically created to replicate in some ways a human driver, right? The human eyes, the human decision-making process. What you notice pretty quickly is that these aren't normal vehicles. It's been equipped with an elegant array of sensors and software. The vehicles rely on a combination of external sensors, which include cameras, a vision system, basically, a radar, and something called LIDAR, which is basically a radar that uses lasers to figure out how far things are and create a 3D rendering.

5:07It can see up to three football fields in any direction and can detect sirens from emergency vehicles, which allows the Waymo driver to visualize the world around it in milliseconds to a high degree of accuracy. Many of these companies have spent years, if not decades, mapping out various parts of the country, the U.S. in particular, as well as machine learning. And the machine learning aspect of it is the part that kind of is intended to replicate our decision-making process. The Waymo driver handles the task of driving in three basic steps. sense, solve, and go. The camera and the vision system sees a pedestrian about to cross the road.

5:49Yeah. They clock that. They figure out how far that person is. The map says, okay, you can go left. You can go right. You can go straight. And the machine learning comes in to say, okay, here are the rules of driving. You know, what, what are your options? Can you just keep driving even though there's a pedestrian? No, here's what you should do. And so all of these things work much faster than how I described it, because obviously driving is a split second decision. But they work together to replicate a human driver in the hopes that eventually this vehicle will drive better than us. And you talked about how self-driving cars have been around in one form or another, or the race at least began 10 years ago.

6:31And why do you think Waymo seems to, at least for now, have won the battle? I think it's a lot of things. I think number one, it's certainly money. They have a ton of money backing them. The other thing is Waymo started off as this research project within Google back in 2009. You know, they have been working on this technology for a really, really long time. And it wasn't spun out as its own company until, I want to say, 2016. Google has announced that what it formerly called its self-driving car project is now an independent company called Waymo. In that time that they were just a research project, all they were doing was mapping and testing.

7:11And it's not just public road testing, or they have like these like fake cities that they create. It's not just testing along those around those fake cities. They also do hundreds of millions of miles of simulated testing in like a lab. It's actually taken this 10 million miles, over 6 billion miles in simulation to get to the answer. Yes, we firmly believe that self-driving technology can help make the roads safer. And they just have always had this advantage over all of these other companies that kind of popped up when self-driving first started actually taking off and when VC money started pouring into this self-driving race.

8:00Gwyn Topham, you are The Guardian's transport correspondent. Thank you for coming into the studio today. Thanks for having me in. So you find me fresh from my holiday to Texas, still audibly high on my rides in Waymo driverless taxis. And I've already been accused in the office of being a shill for the ghost taxi industry. So hopefully you can bring me down to earth a little bit when we have this discussion about what it means that Waymo are coming to London. And can you just start by telling us why London? Why have they chosen our capital city as their first European city? Well, it's a good question.

8:34In some ways, people say there's the very worst place to bring a Waymo. But I think all these tech companies, they want to come into Europe. London's the biggest market for the likes of Uber, for example. And also now the government is very keen to get US tech companies over here. So this announcement came out of the blue to a lot of us that the pilot was going to start next year. why is the government so keen on allowing driverless taxis here? It has been quite a long-term ambition. Look, no hands. And no driver, for that matter. Scientists assure us that it's the shape of things to come in highway travel.

9:10This robot car is the latest project understudied by the Road Research Laboratory. I'm sort of old enough to remember the days when people like Chris Graylin were going, driverless taxis will be on our streets in two years, and even people like Addison Lee, I don't know if you remember them. Oh, yes. They promised robo-taxis by 2021. Private taxi firm Addison Lee wants to add self-driving cars to its fleet within the next three years. Which I eagerly wrote up and get very excited about, so I'm slightly cautious. But I think now the government has finally brought through this legislation by 2027, the Automated Vehicles Act.

9:43What they've done in the meantime is automated passenger services trials, which are the pilot scheme under which Waymo will kick off. I'm in a car right now that is driving itself through the streets of central London. But for years there's been this idea that driverless vehicles could spell jobs and economic growth. There's some stats flying around the 42 billion and 38 ,000 jobs. What, they're saying 42 billion pounds worth of economic growth just from driverless cars? That sounds a bit ambitious, doesn't it? And what they're claiming that the technology will ultimately create jobs rather than take jobs away from taxi drivers.

10:22Well, it's a bit of a specious one, I think. When I look into those stats, they come from a report back in 2020, back pre-COVID, pre-everything. But I think that's where the figures have been plucked from. Yeah, in the short term, it looks like it takes away drivers' jobs. And Britain has a lot of drivers, maybe a million people. And it is possibly less of a job creation scheme than some people might hope. and it's hard to know. Waymo have got software programmers and engineers and technicians all over the world and I'm sure at that level there are people who are very well paid in that kind of industry but maybe the kind of jobs we'll see in London might be more maintenance and cleaning.

11:03When I was cruising around downtown Austin which is built on a grid system it's quite a predictable road structure, very wide streets. And then I came back and heard that Waymo was going to London. I was thinking, oh goodness me, this is a different kettle of fish. Our medieval streets are really kind of unfathomable street planning or lack thereof. How is it going to work? Well, they're surprisingly confident, you might say, but they've done a lot of mapping. They say all the cities are different. They test them in different places, including places that are through the winter, snowy cities, rain.

11:38And I think they know it's a different thing, but what they talk about is not so much just the geography being different, but lots of conditions being different. And every time they do something new, they're mapping something new or trying something new. The Waymo driver crunches all that data and calculates a safe route while responding in real time to the evolving traffic environment. The Licensed Taxi Drivers Association represents all the black cabs. the general secretary steve mcnamara who's a great wordsmith very good on you know it's a fairground ride it'll never work you know this is london it's not for the likes of waymo but i think what waymo feel is actually for all the the things we see as london as terrible driving and jostling in and out of traffic and the impossibility of ever edging out yeah they say it's a very safe city because actually london's got speed limits it's got roundabouts it hasn't got massive intersections where, you know, some tequila young American driver in his SUV is going to bang through the traffic lights or no traffic lights at speed.

12:39So I think this is a place with bumps but not fatalities and I think Waymo is up for that challenge. Yeah, but we are a country that doesn't necessarily respect pedestrian crossings. Jaywalking is not illegal in this country. So aren't they going to have more people just stepping out in front of the cars? They are. One of the things people have mentioned is late night Soho, the impossibility of getting a Waymo down there. But I think possibly Waymo are very conscious of these things and they might end up rolling this out in parts of London that we aren't necessarily expecting. So rather than flooding central London, they might end up sticking it in, you know, zone three, four parts of London where it is a bit more predictable and where those services can operate.

13:22And the Waymos that I took in Austin were spotlessly clean. but I do wonder when there's no taxi driver to tell you to take your rubbish away with you or to stop misbehaving dare I say people have sex in the back of taxis sometimes don't they like what's to stop people doing that in a Waymo when there's no driver to disapprove of your behavior well without speculating too much on everyone else's sexual practices one of the things I'm aware of is that the Waymo will stop or put on its cameras if someone's seat belt is unplugged so certainly it will limit your range.

14:05You are a resident transport expert. Do you think it's going to work in London? I think it could work. I don't really know what Waymo's ultimate business plan is. If it's takeover from Uber and eventually the tech becomes cheaper than the driver perhaps they will say, right, we are, you know with the new uber and perhaps that works but then we have so many different forms of transport that is you know we have like e-scooters that have been on trial for eight years we have the line bikes everywhere we have all these things and maybe they all find their little niche so the pilot goes on for about a year and do you think it's likely that if that's a success we'll see driverless taxis in other uk cities in other european cities yeah i think absolutely certainly Waymo see this as a sort of jumping off point for Europe and actually some people already say it makes more sense in other British cities than in London.

14:58Places that don't have great public transport, places where there's a lot more cars on the road. I mean Manchester for example lots of people do drive around much more than in London I think, only on the outskirts a lot of commuting. So that kind of thing yeah I think they certainly are looking elsewhere as well.

15:20Johanna, when I was in the back of a Waymo, I felt two things. Firstly, thrill, being in a car with a ghost driving. But secondly, oh God, am I safe here? How safe are these vehicles? Not just Waymos, but driverless cars generally. It's a really interesting question because I think it depends on what you mean or what your sort of standard of safety is. the bar that the industry set for itself is they want to create a technology that is safer than a human driver. We have not gotten to that point. So even the humans are notoriously bad drivers. Yeah. I mean, they drink alcohol and they fall asleep.

15:57I live in New York. We were pretty aggressive over here. So yeah, I mean, I can't remember the last time I saw someone actually stop at a stop sign in New York. So even in that case, because the difference is that self-driving technology can't anticipate every single possible scenario that they might encounter on the road. And because it's machine learning and because it's human input and all, you know, there's only a finite number of scenarios that they've encountered. So only a finite number of scenarios that the vehicle is able to navigate to this day. The company argues its own safety data shows 91 percent fewer serious injury crashes and 92 percent fewer crashes injuring a pedestrian compared to a human driver, but acknowledges the vehicles are still learning.

16:44Whereas a human, I haven't encountered every single driving scenario that will ever exist, but we have deduction skills and we can act according to what we see. And so that's still where humans trump self-driving vehicles. And so there's, you know, a world many, many years down the road where maybe a self-driving car will be safer than a human driver or a better driver than a human. But will it be absolutely safe in all situations, scenarios? No. And I don't know that that is possible. More Waymo woes out of Atlanta, but this time it could have ended horribly. The National Highway Traffic Safety Administration in the U.S.

17:27started a probe into Waymo over one of the vehicles inability to yield to a school bus as a way just drove around the school bus and that's an extremely common thing i don't know how school buses look in the uk but in but they're bright yellow and they have when they stop they put at this big massive red stop sign up and there's flashing red lights and the waymo vehicle initially stopped but then went around the school bus. It's gotten the attention now of federal regulators who have launched an investigation to try to figure out why it happened and if it's happened before. And it's, which is the federal agency that handles traffic safety and self-driving, basically said in several statements to various publications that they came away from their conversations with Waymo, basically with the impression that this incident might continue to happen.

18:22So it's not a one-off. Waymo says its vehicle could not from the angle where it was positioned see that the stop sign was out and those lights were flashing. So it did, in fact, proceed to the safety of this technology really is dependent on whether and how the company responds to these really critical safety issues. and much like with the rest of the tech industry or anything that the tech industry does is are they incentivized to protect the consumer or are they incentivized to pad their bottom line, you know?

19:00Coming up, will we one day have cities with no human drivers at all?

19:21The advantages of self-driving cars are obvious. Self-driving cars don't get drunk and get behind the wheel. They're not distracted. They're not tired. They don't need to park up, although I guess they do need to go and recharge their batteries at some point. But what would you say are the downsides to this technology? I think just from a labor perspective, I've heard a lot of concern in the U.S. In California, I've been to several regulatory meetings where drivers are like, this is going to eat into my market share. This is going to take food off the table for me. There is that disadvantage. There are studies that say Lyft, the little brother or little sister of Uber in the U.S., People are taking Waymo more than they're taking Lyft in the zone that Waymo operates in San Francisco.

20:12So this like threat and this fear of potentially losing income because of self-driving is not unrealistic. I think the other part of it is it's really hard to guarantee safety. The safest self-driving technology will ever be, it will be when it's only self-driving cars on the road. Because they can interact with each other and they can very easily communicate with each other. They're predictable, aren't they? Yeah. Until that point, right, it's going to be unpredictable humans with vehicles trying to understand that unpredictable behaviour, as well as unpredictable pedestrians, right? So until and unless we get to a point where it's only self-driving cars on the road, there's always going to be some variable that a self-driving car isn't able to navigate.

21:00And you've written before about this kind of strange period for cities. It's a time of adjustment. I think you've called them the teenage years as driverless cars are first introduced. How do you think our cities are going to look in the future if these vehicles become really commonplace, if they're going to accommodate them properly? It's really funny to be talking about this in 2025, because I remember when we've all of these companies first started talking about the future of self-driving. They were envisioning these cities that were entirely rebuilt around self-driving, you know, because self-driving was the future and nobody would want to drive anymore once there were self-driving cars.

21:40And so we'd have to create separate lanes and we wouldn't need parking structures anymore. And we wouldn't need all of these things that the cities wouldn't be car centric anymore. And I do remember at the time kind of being enamored by the idea, you know, just being like, yeah, I don't we don't need parking garages in the cities like there could be more green space. Yeah, if there was no on street parking, no multi stories. And it just was this vision of the future that I really wanted to believe in by. But now it's 2025. It's actually the year that many of these car companies and self-driving car technology companies predicted there might be self-driving cars everywhere.

22:18And I've seen the reception even in places like San Francisco, which is used to this type of technology kind of driving around the road. It's surprisingly not as accepting and as resounding as I thought it might be. And so it's hard to know for sure. But I think in the next 10 years, our cities will likely largely operate the same way that they do today, whereas self-driving vehicles are on the road. But I think they'll still be driving alongside human drivers and Uber drivers and taxi drivers. It won't be in the next 10 years the only vehicle on the road. And finally, given that you're the expert on this, have you got any advice for our London listeners when Waymo's come to town next year?

23:04Oh, I don't know. I mean, I think give it a try. It is like very, at the very least, like a very fun novelty type of technology and a cool experience. I also would not play chicken with it. I don't test it. No one is learning. Yeah, especially on in European streets. I mean, Waymo is testing its vehicles in New York City. And I'm like, people do not follow the rules here. And I'm like, I just don't feel like they're going to do super well. And so I can't imagine how they're going to do in London. but I would give them some time to figure out the roads before we test out the limits of their technology.

23:46Brilliant. Johanna, thank you so much. Thank you.

23:54That was Johanna Buya. Thank you to her and to Gwen Topham. You can read both of their reporting on driverless cars at theguardian.com. And just before we go, if you're looking for something to listen to next, then Science Weekly, our sister podcast, is serving up a special brew of science, true crime and tea. Ten years ago, while writing a quirky story about the nation's favourite drink, science correspondent Nicola Davis met Tam O 'Baran, a dapper entrepreneur who was selling the dream of a new industry in Scotland. His award-winning homegrown teas were for sale in some of the fanciest hotels and shops in the UK.

24:29And he was helping others set up their own plantations. The story had left a strange taste in her mouth, and now she finally knows why. Join Nicola for a Scottish tea mystery. Just search for Science Weekly wherever you're listening to this podcast. And that is all for today. This episode was produced by George McDonagh and Sinead Campbell, and presented by me, Helen Pitt. Sound design was by Rudy Zagadlow, and the executive producer was Sammy Kent. We'll be back tomorrow.