Self Driving Car Expert: These Are The Only Driving Jobs That Will Remain In 2030! Dr. Jack Stilgoe
Road To Success with Benedict Fowler
Self Driving Car Expert: These Are The Only Driving Jobs That Will Remain In 2030! Dr. Jack Stilgoe
Annotations will appear as you listen
0:00
84:37
Concept
self-driving cars
Self-driving cars are vehicles that can drive themselves without anyone controlling them. They use special technology to see the road and make decisions, just like a human driver would.
Company
Waymo
Waymo is a company that makes cars that can drive themselves without a human driver. They test these cars in places like Los Angeles to see how well they work.
Car
Tesla Model
Car
Tesla Model 3
The Tesla Model 3 is a popular electric car that can drive itself to some extent. It's known for being fast and having a lot of cool technology.
Concept
autonomous driving
Autonomous driving means that a car can drive itself without needing a person to steer or control it. This technology is still being developed and tested.
Term
full self-driving
Full self-driving means that a car can drive itself without any help from a person. Tesla claims to have this technology, but it's still being improved and isn't perfect yet.
Term
autopilot
Autopilot is a feature in Tesla cars that helps with driving by controlling the steering and speed. But it's important to know that it still needs a driver to pay attention and be ready to take over at any time.
Car
Rolls-Royce Phantom
The Rolls-Royce Phantom is a very fancy car that many people think of as the best of the best. It's super comfortable and has a lot of cool features, which is why people often talk about it when discussing luxury cars.
Term
phantom braking
Phantom braking happens when a car's automatic braking system thinks it sees something in the way, like a shadow, and suddenly brakes even though there's nothing there. This can be surprising and potentially dangerous.
Car
Sprinter
The Sprinter is a large van made by Mercedes-Benz. It's used for transporting goods and people and has features that help keep drivers safe.
Term
anti-braking feature
An anti-braking feature is a safety system in cars that can automatically apply the brakes if it thinks there's a risk of a crash. This helps prevent accidents.
Car
Tesla Cybertruck
The Tesla Cybertruck is a unique electric truck that looks very different from regular trucks. It can drive itself with special technology.
Car
Chevrolet Corvette
The Chevrolet Corvette is a fast sports car that many people love for its style and speed. It's one of the most famous American cars.
Term
NHTSA
NHTSA is a U.S. government agency that makes sure cars are safe to drive. They check for problems and set rules for car safety.
Concept
vehicle safety regulations
Vehicle safety regulations are rules that cars must follow to make sure they are safe to drive. Different countries have different rules about what makes a car safe.
Term
driver-assistance technology
Driver-assistance technology includes features in cars that help drivers stay safe and make driving easier, like keeping the car in its lane or automatically slowing down if there's an obstacle.
Car
Tesla
Tesla is a brand of electric cars that can drive themselves to some extent. They are known for being high-tech and environmentally friendly.
Concept
connected cars
Connected cars are cars that can connect to the internet and talk to other cars or systems. This helps them get information about traffic and can make driving safer.
Term
digital mapping
Digital mapping is when computers create maps that help cars know where they are and where to go. It's important for cars that can drive themselves.
Term
steering wheel
The steering wheel is what you hold to turn the car left or right. It helps you control where the car goes when you're driving.
Term
automated speed limits
Automated speed limits are speed limits that can change automatically based on things like traffic or weather. This helps keep drivers safe by making sure they go the right speed for the conditions.
Concept
automotive safety technology
Automotive safety technology includes things in cars that help keep people safe, like seat belts and airbags. These features are designed to protect you if there's an accident.
Car
Jaguar F Type
The Jaguar F-Type Coupe is a sleek sports car that looks really cool and goes fast. It's designed for people who love to drive and want a car that stands out, but it might not have a lot of space for luggage.
Term
V8 engine
A V8 engine is a type of car engine with eight cylinders that helps the car go faster and perform better. It's commonly found in powerful cars and trucks.
Term
Robotaxi
A Robotaxi is a taxi that drives itself without a human driver. It uses technology to pick you up and take you where you want to go, just like a regular taxi.
LIVE
By 2035, every driving job in the UK
gone. Delivery drivers, taxi drivers, careers. How realistic is that scenario?
So I would say Dr. Jack Stillgo.
Dr. Jack Stillgo is a professor in science and technology studies at University College London
where he researches the governance of emerging technologies.
You've warned that automation is arriving faster than people realise and that the UK isn't ready
for what happens when machines start taking the wheel.
You know, when it comes to artificial intelligence, a lot is happening in a lot of areas very quickly.
Automation and self-driving vehicles could trigger the collapse of one of the biggest industries
in the UK. Autopilot does go wrong. The first person to die in a Tesla autopilot crash
was a real Tesla enthusiast. When something goes wrong, who's responsible? Is it the driver,
the developer or no one? So you mentioned that you could take your hands off the steering wheel
on a motorway if you own a Tesla. Please don't do that because...
Dr. Jack Stillgo, your work on how society governs technology puts you right at the intersection
of innovation and consequence, asking not can we build it but should we be doing this?
You've warned that automation is arriving faster than people realise and that the UK
isn't ready for what happens when machines start taking the wheel.
But in your own words, who are you and what do you do?
Thanks very much for the introduction, Ben. I'm a professor of science technology policy.
I work at University College London. The day job, I teach students who come to our department,
come to our degree programme with interests in how science and technology is changing the world.
So I work in a university but I spend quite a lot of my time working with various bits of
the British government because I want to help them understand how you come up with good policies
in this area. As we all know, when it comes to artificial intelligence, a lot is happening in
a lot of areas very quickly and there's a sense sometimes that we're sort of suffering from a
sort of whiplash and what do you do about that is not just a question for the scientists and the
engineers, it's a question for society. So let's imagine a headline for a second.
By 2035, every driving job in the UK gone, delivery drivers, taxi drivers, couriers gone,
replaced by fleets of self-driving vehicles that never get tired, never take breaks and never
make mistakes or so be it we think. So it's just a theory. How realistic is that scenario in that
headline? So I would say 2035, never going to happen, right? I mean it's just one of the things
that I'm interested in is asking what it would actually take for scenarios like that to happen.
So if you are a technological enthusiast, you might be a startup guy running a self-driving car
company or you're fixated on artificial intelligence, right? There's a lot of hype there and a lot of
money behind that hype that says everything is going to change instantly. Something like cars
though, streets. Streets are really, really complicated things. The world does not change
that quickly from one form of technology to another form of technology, which means that
we are likely to have a mix of things on our roads and this isn't just in Britain, this is going to
be everywhere in the world, a mix of things on our roads for the foreseeable future. I mean,
genuinely for hundreds of years. So there are changes that happen quickly in terms of the
ways that people get about and we can look back, say, 120 years to sort of American cities when
you saw big, quick transformations from trams and buses to the arrival of the motor car, right?
But you still had horses on the streets and one of the things here in London that you still
have on the streets in 2025 is horses, right? You don't just switch from one thing to another.
So anytime there's a new form of automation, it has to deal with the world as it is,
which means a really, really complex engineering problem.
See, I spend a lot of my time also in America, specifically in Los Angeles,
and I've been starting to use Waymo and a bunch of self-driving cars that are in LA and I remember
my first ever drive I took in a Waymo. I think everyone probably would because it was such a
unusual thing to be doing, so out of habit, so out of control. But after using it over and over
again, I actually prefer it to the other methods of getting around Los Angeles.
In the UK, we're not currently there in 2025 yet. We're at the point where we can take our hands
off the wheel on the motorway in something like a Tesla Model 3 and have a degree of self-driving
cars. But when do you think like the launch date, like what you were saying about horses kind of
are still here, but the car came about 120 years ago, when's the launch date of like
full auto-autonomous driving? So I think the boring answer to the when question
is that it's the wrong question, right? It's going to be a where question. So where will
self-driving cars appear? Where will they get used? Under what circumstances? There's a reason why,
and I want to ask you about this because I've actually never been in a Waymo without a driver,
and that's new to me. There's a reason why those things are in only some parts of America,
right? And there's a reason why we don't have Waymos in the centre of London without drivers.
We do have some forms of, we have some small self-driving car companies testing their technologies
here, and that's because places are really, really different, right? Just because your robots
learn to drive in Los Angeles doesn't mean it can drive in San Francisco, doesn't mean it can
drive in Phoenix. So Waymo has had to work out how to deal with all of those different cities,
which means that the when question, it sort of depends because the companies that want to
deliver these technologies, they will have particular ideas about whether they think
there's a market there, but also particular ideas about whether they think it's an easy thing to
automate. And as anybody that's been in Los Angeles knows, the streets are so different from
the streets in central London, so it's actually a really different engineering problem. Just because
you have, as it were, learnt to drive in the United States doesn't mean that you can then just pick
up your self-driving car, put it on a boat, take it to London, and then also consider
properly, properly mad driving around and take it to Naples, take it to Delhi, right? And that's a
completely different engineering challenge, plus those companies are going to think, well,
are we going to get the same sort of market in those places that we're going to get in Los
Angeles? Probably not. Knowing what you know about self-driving vehicles, would you have got in that
in LA? So yes, I would, because I know Waymo quite well, and I've been really impressed with how
they've developed their technology, and they've done it in quite a responsible way. To go back to
your Tesla example, though, I think Tesla is maybe a more real example. So you mentioned that you
could take your hands off the steering wheel on a British motorway if you own a Tesla, and I would
say please, please don't do that, because I think legally you're not allowed to, but also the Tesla
technology isn't nearly as good as Tesla say it is. So they've marketed the technology for more
than a decade as an autopilot, and more recently been calling it full self-driving, and it is not
a full self-driving technology, which means that it does go wrong, and when it does go wrong, we
don't know very much about why it's gone wrong, which means that if you are experimenting with this
stuff behind the wheel of your Tesla, you are taking your life into your hands. We haven't had any big
crashes as far as I know in this country, but in the United States, there have been, you know,
quite a few, more than a dozen serious incidents, deaths caused by people using that technology
in a way that is really, really dangerous. So if you'd asked me the question about the Tesla
RoboTaxi that's just being rolled out in, I think it's Austin in Texas, I might have had a different
answer. Waymo, I've been really impressed with because they've built a culture of engineering
carefully, developing the technology carefully on the whole, getting on quite well with the
governments of the places where they're operating. They seem pretty trustworthy.
So how then, if what you're saying and your viewpoint on autonomous driving in Teslas,
and the fact there was a word there like illegal for some of this, how then is it actually being
used in the UK? So I don't know the legal status of the full self, the most advanced form of the
Tesla technology. I know that there are a lot of Tesla users, Tesla customers using that technology
in the States and Tesla use their customers experiences as training data for their artificial
intelligence systems. And I know that one of the frustrations that Tesla have is that
their full most advanced technology is not fully legal in European markets or the UK,
which means that like a lot of this, we're in a sort of legal gray area. And it means that it
sort of falls on users to decide whether or not to do this stuff. And the bottom line is that as
the driver, you are still fully responsible all the time for what your car does, which means that
you need to really know the limits of the technology. And at the moment, it's really hard
to know the limits of the technology because we still don't know when these things fail. We've
had incidents in the United States, for example, of something that's called phantom braking,
because Tesla, as part of their AI technology, use cameras as their main sensor, which, you know,
they've got very good cameras and Tesla's argument is, well, we as human beings only have these
sensors, which are in effect cameras, and we can drive. So why not just just use that? But it means
that occasionally the technology gets fooled by things like shadows and so can slam on the brakes.
And there was an incident a couple of years ago in the United States where a Tesla slammed on
the brakes in a tunnel. And I think thankfully nobody died in that incident, but there was big
pile of cars behind it, right? Lots of cars slamming into the basket. It's funny, you say that my
Sprinter in the US has like an anti braking feature, obviously, where if it sees somebody
come out in front of you or he thinks you're distracted, it will just break for you. It's got
all the kind of systems that Merck's fit on it. I had a weird thing once where I was going under
a bridge with a massive shadow and it just locked on the brake in the middle of a freeway. So it
might not be a Tesla specific, but Tesla have obviously taken that technology to the next level.
And there's a quote, which I heard on another podcast, and I really do like it, which is,
some people want to go to Mars and others want to take over the universe.
And what it was referring to is the difference between Elon Musk and Sam Altman at Open AI.
Do you think Elon cares more with his philosophy of his businesses and the world and his holistic
approach and what he actually wants, which is to take humans to Mars than a lot of the recent
exploding of startups around AI and putting that technology into things like vehicles?
Or do you think it's dangerous? I mean, I have no idea. Trying to work out what Elon Musk actually
wants is really, really hard. It appears to change from time to time. It's not that we
shouldn't try to work out what drives these people and what drives these companies. And I have no
doubt that somebody like Elon Musk is honest about the potential that he sees for a technology
like a self-driving technology to improve people's lives in some ways. And we might
take issue with how fast he wants to go and what role he sees his company play relative to
governments. But I'm not sure I'm any better of a muscologist than anybody else.
I was in a Cybertruck about a year ago and we went from Las Vegas to the Grand Canyon in this
Cybertruck and then we came back the following day. And we had a VA Corvette, was the car I'd
rented and my friends had rented a Cybertruck and went to do this video. And we drove back
in the Cybertruck in full autonomous driving mode. And it was actually insane when we got into
Vegas how one of the things that it does to kind of calm you down, that it knows what it's doing,
is giving you a big picture. So like on the screen, it picks out every individual people and
sort of shows you. And then you can kind of fact check what it sees. However, at the same time,
we had towels wrapped around our head that we'd taken with us to be stayed for the night
because it had an absolute meltdown and was just bonging constantly. Full chat and we just couldn't
get this bonging noise to stop and it's because of something with the door sucking it in.
So you're kind of putting your faith in something driving you at the same time it's having a meltdown.
With what you know and what you regard things and it can't even do the easy things.
With what you understand about self-driving Teslas, would you get in one and drive back
from the Grand Canyon to Las Vegas for four hours? I would not. No. No. I mean, I genuinely
wouldn't. I mean, one of the first things I did when I started my research on self-driving cars
was I was living in the States at the time. I was living in Colorado and I went along to a Tesla
showroom and I got them to and I said, I'm a British person with a posh accent and I'd like to
test drive one of your cars and they went, yeah, fine. And I asked them to demonstrate what was
then called autopilot, which was a less advanced version of the technology. And I wasn't just
interested in how the technology worked or it didn't work because it was really impressive.
The first time you do any of these things, there is a sort of sense of magic that you think it
shouldn't be able to do this, but it seems to be able to do it. I was also really interested though
in the psychology of it, in a human's reaction to being driven because one of the scary things
for people that have maybe been in a waymo or sat behind the wheel of a Tesla is you switch on a
system like this and you think that is amazing and then it instantly becomes boring because if it
works, it's the most boring thing in the world. It's like being driven somewhere. You don't want
it to be interesting. You don't want it to feel like a thrill. Otherwise, you know, something's
gone wrong. And that's a well-known thing within social science is a problem that we call automation
complacency is when something purports to be automated, so an airplane autopilot maybe,
what we know happens is that people who are supposed to be in control of these systems tend
to switch off. And we as the human being do that very, very quickly. So we go, okay, this is working,
which means that if there is something that goes wrong with the technology that we don't
understand, we are very, very bad at doing what we need to do in order to safely take control
of that technology. And the sort of the switch where it's either the machine in charge or it's
the driver in charge and switching between those two things, we know that that's a really
problematic thing to do. Now, Tesla are not good at acknowledging how complicated that is,
which is why they promise that their car in effect can just drive you from one place to the other.
You say that you understand slash work with the government on these things in various areas.
How are these being allowed on the road in their current form as they are?
So in the United States, there are a lot of experts who object to them being allowed on the
road in their current form, particularly Tesla. We have just seen the regulators in the United
States, so NHTSA, they're called the National Highway Transportation Safety Authority, I think
is the acronym, are doing investigations into Tesla's technology. The United States has a
different approach, which is that if you are a company and you want to deploy some of this
technology, you can a lot of the time just do it and you can in effect take responsibility
for it yourself, the technology. And if something goes wrong, then it plays out through the courts.
In Europe, the approach is rather different, which is that you need to persuade regulators
that your technology is safe and works pretty well before you're allowed to put it on the road.
So we have all of these systems to ensure that when a regular car crashes, it crashes safely and
doesn't crush the legs of a driver and all the rest of that stuff. And so we have regulatory
systems that test these things beforehand. Do you think when the kind of normal person
with the knowledge of self-driving cars or new technology, like if I take like a
lined average of how interested the average human is in this sort of stuff,
here's that something has passed regulation to be allowed. I think we have this automatic sense of
it must be okay then. I think that's part of the worry, right? So when you're asking me if I would
get in a sort of self-driving car, one of the things that I'm doing is the same calculation that I
would do when I'm getting on board a plane, right? I don't want to have to walk around the plane and
say, does this thing look safe, right? As the passenger, I'm thinking, has somebody who knows
more about this stuff than me being around the plane, checked that it's safe and you have all
of these regulators in place and airlines that you know are going to take responsibility and you
think, yeah, I sort of can trust them and that's why. And I know that aeroplanes have very, very
few crashes, so I'm going to trust them and get on to the plane without really having to think
about it. And in the case of a self-driving car, you know, asking a user to make that calculation
is sort of unreasonable, I think, which is why it's so important for regulators to be on top of
this stuff. Because even, and actually Elon Musk has acknowledged this point, that Elon Musk has
acknowledged that the people that he's most worried about are the people that consider themselves,
you know, Tesla experts, because they're the people that might be most tempted to try out
their autopilot technology. I now feel like I'm slightly accusing you of doing this on your way
back from Las Vegas. But it's the people, you know, it's those people who think, I know about cars
and I'm going to switch this thing on and I know that if it goes wrong, I'll be able to take over.
And there's a sort of complacency that means that those people might need to be the people
that you worry about. It was really noticeable that the first person to die in a Tesla autopilot
crash was a real Tesla enthusiast. He was a YouTuber with a Tesla channel, a guy called Joshua Brown,
who back in 2016 was hurtling down a divided highway in Florida. And, well, we don't know
what happened because he died instantly. So all we have is the data from the car afterwards.
But the car passed straight underneath one of these giant American trucks,
passed straight under its trailer, crushed the top of the car, killed him instantly. And he was,
as far as we know, sort of doing an experiment with the technology, which wasn't reliable at the time.
So it begs the question, when something goes wrong, who's responsible? Is it the driver,
the developer or no one? And does that kind of vary country to country?
Yeah, well, that's one of the urgent things that needs working out. In Britain, we've just been
through a process of trying to get a bit of clarity about the rules for that. So about a year ago,
government passed the Autonomous Vehicles Act, an act of parliament that tried to clarify
under what circumstances different parts of the system would be responsible. So would it be the
driver? Would it be the person that had made the self-driving vehicle? Would it be the company,
say, operating the self-driving vehicle if it was being operated by an Uber-type fleet of cars?
Would it be the insurer? Or would it be something else?
There isn't one answer to the who's responsible in every situation. It will vary just as it does
whenever one car crashes into another car and the drivers have to hash it out with
insurance companies and the rest of it. Who do you believe should take moral accountability?
So it's not clear a priori who should be held morally liable, but what I do know is that when
it comes to new technologies, what often happens is that the people who are selling the new technology
want to evade that technology. Sorry, they want to evade responsibility for the technology,
which means that we need to be more precise about questions of responsibility when we have
new technologies. So the easy thing to do, so to go back to that Tesla example I was just mentioning,
when Joshua Brown got killed, the easy thing for Tesla to do in those circumstances and
circumstances like that is to say, well, it's the driver's fault, right? Because when you get into
a Tesla, you are supposed to have read all the small print about what your system can and can't
do and the small print will say legally you are responsible for this technology. But it's
more complicated than that because Tesla are also lulling you into thinking that this car can drive
itself and especially if they're selling a technology called full self-driving, you can be
forgiven for thinking, well, this is a technology that can drive itself. In that circumstance,
it was a case of a self-driving car crashing into a non-self-driving vehicle, which would have been
a traditional lorry. But I have seen firsthand when I've been taking Waymo's and I think I've
taken enough of them now to get a real sense of kind of what they're about. Yeah, they're like
that Waymo's get into the most problems when they're dealing with other Waymo's.
I think early on, there was a few instances where like if you come up to a crossroads with
two Waymo's, neither one of them knew which one to kind of go first. But what it does to me is it
begs the question, which is if a self-driving car crashes into another self-driving car and someone
dies, would that then just be AI battling AI? Because if the answers are so
beyond our moral compass because it's technology dealing with technology,
would we then use kind of AI to fight that battle of who's responsible?
I haven't thought about that possibility for the simple reason that there haven't been enough
self-driving cars on the road for that to be a thing that we've yet had to worry about.
Part of the issue there, the question of moral responsibility is complicated by
the fact that the technological systems are quite often opaque. So we don't know why an AI did what
it did in quite a lot of circumstances. So for anybody, like with any I'm imagining in that
example, you would have an insurance company, maybe lawyers, maybe a court if it goes to court
trying to work these things out and they'd be trying to work out who was responsible
in order to assign blame and well in America deal with amounts of money and things.
The bigger question though for society is how we can learn from things like that so that we
improve. So society's interest should be how can we make things safer over time and what we got very
good at with aeroplanes was learning from crashes. So when an aeroplane crashes, yes we want to know
well who was to blame but in lots of cases regulators will say actually we're partly
interested in who's but we're interested in how to make sure that this never happens again
and that process of learning from accidents is why we have accident investigators and all
the rest of it. So one of the other things that we're currently trying to work out with the government
is about what the requirement should be for accident investigation when it comes to something
like self-driving cars. How much for example should the police be able to take in terms of data from
a self-driving car to say well you know so that we can exactly trace the movements of this vehicle
but also trace if you like the sort of thought patterns of this vehicle to work out why it
did what it did so that we can stop that from from happening again. Many of you might not know
this but away from the recordings that I do in my van studios I've actually got a digital
marketing agency now we specialize in a lot of automotive clients but we cover everything
really. Our team is made up of PPC specialists SEO specialists and the most talented designers
I've ever seen which have done work like the Starnegloss website the TWR website and many
more. We've actually just built icon box for the auto alex crew as well meaning that people that
watch their channel can buy their favorite merch seamlessly and in style so if you're
interested in starting a project and you'd love to speak to us just tap the link below and let's hop
for our conversation with the belief obviously that can be changed throughout our conversation
but with the belief that I believe that a self-driving vehicle is safer at the minute than
a right than a human within the vehicle and the reason that I have came to that conclusion is
actually based on a recent story that I've had to deal with in my own life which was a couple of
weeks ago I live about an hour and a half away from Gatwick airport and my other half and her
friend flew back in at one o'clock in the morning now Toby who's behind the camera tonight who
takes a lot of drives for me and I trust him implicitly because I've seen that he doesn't
fall asleep he's just awake 24 seven like monster energy and red bull powers his life
wasn't available to go and pick them up I don't drive at that time of night because I'm absolutely
useless so I deem myself as unsafe so I just know I'll have my head out the window in the cold
eventually until I have to fall asleep at Redding services yeah so I called my local taxi driver
that I knew and as airport runs to book a ride to pick them up collect them and take them back
in the belief that he would be professional safe and right to do so I then told both my girlfriend
and her friend that I trusted this man and that he'd done right for us since I was little
and I went to bed that night believing that everything was going to be okay yeah and it
wasn't he nearly crashed on the m40 from being half asleep he then nearly well he did he mounted
the verge went up the verge and they caught it on camera several times and they were so petrified
they didn't know what to do and what to say in the back of the car they also then woke up because
they were falling asleep when he couldn't find one of their addresses and it was late at night
very dark in the middle of the country he's going back towards the Cotswolds and they woke up on
a random farm and thought he was kidnapping them and you had to calm them down for a second when
my partner got in hours later in the morning after he was doing 25 mile an hour down the motorway
I was livid like like beyond angry and I just couldn't help but think in the days that followed
that if a Waymo had existed at Gatwick and I could have booked it to drive back
those two to where I live I'd have done it yeah because based off my experience in them
I believe that that vehicle won't get tired it won't lapse in concentration there's nobody
in it with any moral compass to do anything wrong and therefore you can't technically
get scared as your passengers within it so you make it feel safe and sound but
would you have still picked the taxi driver over a Tesla Robo taxi
so yeah interesting what I mean obviously not knowing what I now know about that taxi driver
right or even that you know if it was a mini cab company that mini cab company which has
betrayed your trust right your trust in what you thought was a safe system involving a
human being and that's why we have rules about who isn't who isn't allowed to drive and if you're
a professional you know the circumstances in which you're allowed to drive and you know having breaks
and not driving tired and all the and all the rest of that and in in that version you know there
is a sense in which profoundly you have been let down by that person who has broken those those
rules just to take a step way back though to why people are interested in self-driving vehicles
it's really important to remember the flaws of human driving and this is something that
drivers often aren't willing to admit to themselves but yes humans most of the time
are fine and most of the time there aren't incidents but we also know our deep down know
that we have important flaws and we know that fallibility can be really really dangerous
and we all you know know people who you know will occasionally break those rules whether it's
speeding or drink driving or anything else and that should rightly terrify us plus we have
the stats on how many people die on the year i mean so more than a million people every year die
as a result of the technology of the motor car right which is a brilliant technology that has
enabled all sorts of social benefits but it also has this massive massive risk this giant risk that
you know if for other technologies if you said we're going to you know enable you to do all
these brilliant things and go and travel places and work in different places but
a million more than a million people a year will die of this thing right so if you're an
engineer you're thinking well how can we solve that problem and that i think is some of the real
excitement behind trying to develop these forms of automation and i can i can totally see that
do you think that number is going to skyrocket in the next 10 years on a graph
i so i i look at some of the trends and there are reasons to be optimistic that
actually a lot of places like britain lots of european countries are getting much safer every
year right so if you're in a car um on the whole you get safer and safer every year because you
technology around you improves and all the rest of it but then you get some some reasons to be
rather depressed about it like the stats in the us are not going down so the us is not on
average getting much safer which is which is rather sad i can probably ask you to guess why
you think the us is getting more dangerous for driving do you think that is down to the self
driving i don't it's definitely not down to anything to do with self driving just because
the numbers aren't big enough yet but it is down to technology at least in part a bit like my van
breaking scenario but it's actually it's it's people it's distracted driving right so that's
the big new risk that's happened the last 10 20 years is people on cell phones which in the us
you know they are more relaxed about than than we are here in here in europe so there are reasons
in terms of what the technology of self driving will do you know to go back to your example about
how you make a calculation between human drivers and and and and robot drivers one of the things
that we should be interested in and one of the things that we should be trying to work out is
yes averages so on average our self driving cars safer than human beings but my sense as a sociologist
my sense is that ordinary people won't just be thinking about that they'll also be thinking about
can i trust that a self driving car is going to be good in this circumstance that i know a human
will be really good at and vice versa so we also want to know under what conditions a self
driving car will fail and one of the weird things is that you know robots and humans do not drive in
the same way so actually robots will fail in a different way from humans like the phantom braking
stuff right humans are not going to get freaked out when a shadow comes from a from a bridge but
if you're relying upon a particular system that deals with those sorts of things and has to make
decisions about about that then that might be a new risk so if i were to book you a vehicle or give
you the choice and booking a vehicle from Gatwick Heathrow airport right across through the centre
of London to the other side with all the challenges it faces and i gave you the opportunity of being
driven by a driver in a tesla that would likely use itself driving finishes versus an uber yeah
which one are you choosing where do you feel safest i'm probably so in both circumstances
i would trust the driver now if you were to take you know put my trust in the driver which also
means putting trust in the driver to use the technologies that they have at their disposal
right so if though waymo or tesla said well we're going to trial our technology and and take a
take a and we think it's good enough and we're not going to have a human behind the the steering
wheel then that would be a different sort of calculation because then i'd be thinking well
who do i trust that this technology has been well tested what do i know about what's going on
behind the scenes of the of the technology which is another thing that we should talk about you
know even though there's nobody behind the wheel there are always people behind the scenes helping
these helping these things out so you would make that trust decision based on what you knew about
about the system in the case of it being you know gatwick to london through london the things that
i would know was that well actually nobody has yet tested their technologies on that route therefore
nobody knows if they can do it well and when you're driving in los angeles when you're in
in a waymo in los angeles one of the things that i'm imagining you think that enables you to trust
the system is you you know that waymo have been shuttling around those streets now for a few years
right they're pretty well known streets so they're that massively reduces than possible surprises
it seems to me from the research i've done that there are several big players attempting to bring
that exact thing to london to the uk as fast as possible one of them i've read is uber equating
by 2027 they'll have their full self driveless cars and taxis available in london to kind of ferry
people around a bit like what you said about elon making statements about stuff that don't
actually come to frish for you later do you think that that is likely or unlikely i think that's
unlikely um so there is the it's really important to realize there is a huge amount of hype um
because you know the companies involved know that they they want to be associated with this new
technology right everybody wants to be at the front and some of them well some of them imagine
that it's a race you know i don't think it is a race in that one person is going to win and then
you know stitch up self driving across the whole world because one of the things that we know about
different places is that forms of transport look very different across those those different places
uber were one of the most interesting companies in this space back in the day so you know in the
sort of 2016 to 2018 period um and one of the reasons i got interested in this stuff in the
first place so after tesla had had a few crashes actually the most prominent and most shocking
crash involving a self-driving car was a test a prototype uh volvo being tested in
arizona by uber so they this this was just a prototype it just had a um a safety driver
behind the wheel um and that uh in 2018 that vehicle ran into and killed a woman who was
pushing a bicycle across one of these big american wide uh wide roads um and uber stopped their self
driving car program after that because that was one of those moments when actually regulators
did wake up and say we don't want our roads to be a laboratory for the testing of technologies
that don't seem to work um and uber it was felt were going too fast too quickly they were doing
irresponsibly they didn't have a good system of you know making sure that if something goes wrong
there's a human there to um to to take over and and and all the rest of it so they've only actually
relatively recently resuscitated their self-driving efforts in collaboration with some of these some
of these other companies this is pro linebacker tj watt and i'm back with ypb by evercrombie for
another active wear drop my second co-design collection has new shorts and tanks that keep up
with all my in-season workouts and their new restore collection is a game changer off the
field too because even pro athletes like me need rest days shop ypb by evercrombie in the app
online and in stores because your personal best is greater than anything
we're in london right now i'm growing up as much as i think about the positives like everyone
does about a place or a city or the things that you want to go and see you know big ben the london
eye the hustle and the bustle the people the cultures everything about it that makes it great
but you also remember the moments that are truly shocking and um terror springs to mind i've just
watched the documentary i was too young at the time about what happened with the bombings on the
tube and the buses which were truly shocking and every now and then we get a reminder of how dangerous
terrorism could be do you see that as a huge risk to cars that are automated driven by a network
so well i think this is one of those areas right where we need to be really clear on not just
is the technology on average say for the human driven cars but what how might those risks change
how might technologies be vulnerable in different ways from human beings and one of the risks that
people especially sort of cyber security people have already identified is that if you have fleets
of connected vehicles you potentially create you know they talk about things like attack surfaces
you take you create new vulnerabilities where somebody could take control and you know bricks
say an entire fleet of vehicles which are all dependent on this on this system all up to 200
mile an hour and off you go yeah or just you know turn them off more boringly and stop people from
being able to get where they need to where they need to go so you introduce new vulnerabilities
that having you know you know properly so the word autonomous is normally associated with
autonomous vehicles but you can talk about human beings having autonomy and one of the
things that we value when it comes to driving is the autonomy the freedom to be able to drive when
and where we want and that's our choice not the choice of you know Tesla or whoever is is is
operating that fleet so yeah we should be worried about the new risks that that these new technological
systems create and it's not just the idea that you've got a car you've plucked out the human
you've dropped in a computer and everything else stays the same right it's not going to be like
that in order to get self-driving working in a safe way you might have to completely reconfigure
your system so that the cars are all connected to each other and to some base in order to you
know make the thing safe i think the normal human beings everyday access to autonomy is probably
chat tbt and i always say wrong chat tbt i've got this the crazy pro one at the minute and i'm playing
around with loads of different things and what i find fascinating is you can kind of see that
consciousness ability of it starting to come through you know the way it speaks to you the
way it connects with you right it's becoming a little bit different and when people refer to
self-driving cars the the argument for ages has been oh what does the car do when the child runs
out in front of the car and there's like a group of guys over there does the car go into the guys
or drive into the child like what would you the car do what would the system do compared to a person
but it's it opens the door to ask a really good question which is how much of an autonomous
vehicle is being led by ai versus kind of systems and processes like you'd program a robot at school
as a kid yeah and it's and and this is one of those areas again where we can we can we can really tie
ourselves in knots about it so you know i think we've probably all had that that weird experience
with something like chat gpt right or a large language model where you think this is eerie this
is magical is there something going on here you know and and and people have already sort of
ascribed sentience to it and and the computer scientists would stamp down on that very quickly
and say no that is you as a human being and doing something that we know happens right which is
anthropomorphizing which is in the same way as you know we talk to our pets it's very easy to assume
that there's more going on there than is actually going on there it's ultimately just you know
predicting the next word and and doing it really really effectively in the case of a self-driving
car one of the things that the companies developing these things have to make decisions about is how
much autonomy they give to the decision-making within the car itself versus how much you program
in rules in itself so in the case of driving you know obviously the rules of the road something
like a speed limit tesla already got into trouble that for example their full self-driving system
sorry full self-driving system put that in quotes because it's definitely not full self-driving
um was making decisions that broke the official rules of the road in the united states like
rolling slowly through stop signs and doing that sort of thing but those are things that
almost everybody almost every human does so there's a really difficult trade-off you know do you
obey the strict rules of the road or do you do what humans do which is to you know recognize
that there are gray areas and lots of times when you'll have to make a decision um when you know
or you do it because everybody else does it right it's more of a norm than a rule what do you think
say fest um i think in the case of an automated technology most people would expect those technologies
to obey the rules of the road more strictly than they would expect humans to so it's much harder
to forgive a self-driving vehicle for breaking the speed limit for example because
we sort of know that a self-driving vehicle can easily just or just just limit itself
but there are some interesting examples like at the moment in in europe we have the introduction of
intelligent speed um what's it called intelligent speed assist or whatever it is but technologies
that can stop you from speeding um which should be really good for public safety but also lots of
people disagree with them because they feel like their freedom is being is being constrained so it's
one of these areas where you know it's complicated and to pretend that it's just an easy thing to do
if you're programming a self-driving vehicle you know because ultimately the sort of central point
if you're an engineer making a self-driving vehicle you are involved in one of the most
interesting technologies out there right now but but you've got to work out all sorts of complicated
things about how to drive through a world that is ultimately being designed for and around human
beings does a regulator understand that technology though i think one of the things that regulators
will have to do in the next decade or so is is is come to understand more about the technology
at the moment i don't think they do um in the case of so some of the crashes that i discussed in the
united states one of the things that's happened is that the crash investigators for those have
had to come to understand the technology really really quickly and i spoke to some of the crash
investigators and it was really interesting getting their sense because these these guys know
everything about you know road safety or air safety and they're the people that
investigate plane crashes and everything else and when it comes to a self-driving car crash and they
go what went on here and we don't know why this thing did what it did so can we look at the
data within the within the car's computer and sometimes that data doesn't reveal everything
that you want to know about about what it does so there might be actually limits on how much we
can understand about about why a technology has made a decision that it's that it's made for
that make this really difficult which is why we need the people developing the technology to
behave more responsibly than they're behaving at the moment over 300 000 uk workers depend on
driving income whether that's logistics you know driving lorries that's driving amazon vans around
automation and self-driving vehicles could trigger the collapse of one of the biggest
industries in the uk and we keep coming back to los angeles and you know lands far far away to
many britains but it's happening there you drive down the road in a waymo and a shopping trolley
that is powered by a battery goes past you to drop off someone's goods yeah or a waymo full of
someone's order that they've ordered and the guy guys come out and put it in the way it's
sensing on its way and you can now have food delivered and a great example of why someone
would choose one over the other as well is i got two quotes to get back to the airport from where
i was in la in pre-knowing that i'd have this conversation at some point and wanting to refer
to it and the uber was 58 dollars sending back to the airport and the waymo was 22 it was like
more than a 50 percent difference between the two so do you think if you if you were working
as a lorry driver amazon delivery driver and considered that to be a long-term
income would you be worried so i mean yes i would be worried i mean there's a you know
some of the things that you talked about just there though are disruptions that have happened to
people's livelihoods that don't have anything to do with with automation or where the forms of
automation are quite so here in london for example 20 years ago if you were a black cab driver you
might have you know quite a nice wage and the big disruption to have come for you is likely to have
been uber right which is not because there's a robot driving the car but because there's
one company that decides that it can do a clever thing involving the market for in effect mini
cabs right massive massive disruption to how people driving make money in in london and who
that is is is opened up to i mean i've already expressed skepticism with the speed of the arrival
of of the technology simply because these environments are quite complicated
ones for robots to to drive around in you know places places like london give us an example
what of a complicated what of why london to someone that might live in australia or new
zealand i know it listens to why london creates a such a challenge compared to la because most
of la just as a starting point is a grid system right right absolutely and so so if you've been
to cities in the united states you might know how roads work that there's often multiple lanes the
junctions are often governed by traffic lights which you know in sort of digital terms are quite
useful on or off technologies in london your you know the first thing to recognize about
the middle of london is that it predates the the road layout predates the arrival of the
motor car by hundreds of years right so the so the the the oldest bits of the city are you know
roman medieval which means that the roads were not designed for cars let alone self-driving cars
it means some of our junctions are complicated we have things like roundabouts that rather than
being on or off from much more sort of gray areas about judgment which are a harder thing for a
self-driving car to have to deal with i could often take my students out of my classroom
and show them different roads around london so my university is quite near euston road which was
one of these a fairly modern road introduced in london in the 1960s which has a particular design
philosophy around it which is that it segregates off pedestrians buses
cycles and cars right so they each have their lane and those forms of transport stay in their
lane literally um there are other bits of london so if you go to there's a road called exhibition
road which is where all the museums are which is something called shared space where the designers
have said actually we're going to blur the lines between the pavement and the and the road
we're going to make the cars go slower because the cars might expect that a pedestrian is going
to walk in front of them at any time right so it's a different design philosophy designed
around human perception really really hard to go through there if you're a self-driving car because
those you've introduced just much more uncertainty into that system i've got a um a roundabout near
me and people living in that local area will probably be commenting on this now because it's
famous but have you ever heard of the magic rather the famous magic roundabout in swinton
yeah and it is it's near me but do you think when we speak about regulation we speak about
true autonomy and self-driving cars we speak about road mapping and you're obviously very
vocal on the fact that london and the uk presents different challenges to america do you think it's
possible for a self-driving car to get through the regulations allow full autonomy of someone like me
to buy five tesla robo taxis and send my fleet out to work every day and crack on with it but
they just wouldn't map that one roundabout because it's just one thing in the uk do you think that's
going to slip through and be where a lot of problems lie so i think it's that sort of thing
though that might yeah it might be where where problems lie right so this is why i think rather
than saying rather than asking when will i have myself driving car right asking where becomes
more powerful because it might well be that actually expecting your tesla to drive all the
way from bristol via swindon to norwich or somewhere you know is going to be impossible because
no company will invest in the sort of in what needs to go on behind the scenes which is things like
digital mapping right it's like making those things um more predictable so that that route
becomes possible but if you want to get a cab from one side of milton keens to the other right and
milton keens is a much more american seeming sort of grid layout car and uh car environment you might
that might well be possible um relatively soon and that's and actually a lot of these companies
are testing in and around milton keens for this reason because it it feels like quite a good sort
of a good laboratory to all my loyal listeners listening on spotify apple and other streaming
platforms i urge you to do me a quick favour that you might not know that you could do you
can actually follow if you're listening on spotify the road to success podcast and also rate it with
how you feel these conversations have been how they may have helped you or if you're just enjoying
the one that you're listening to today it really will help us if we're able to grow our streaming
platforms beyond hundreds of thousands of monthly listeners thank you so much for listening to
this episode of the podcast and i really hope to bring you some more inspirational guests soon
do you think we'll get to a point in your island's lives where 50 percent of the road is buzzing
around without operators in the vehicles so that so this is the other part of the of the challenge
right and i think this is definitely worth considering if the technology is genuinely
beneficial and society agrees and government agrees that yes we really want to encourage
this technology in the same way as you know in london they want to encourage people to cycle for
example so what do they do they don't just tell people to cycle they try to make the environment
better for cycling okay so if you really wanted to support self-driving technology one of the
things that you would do is is to say okay well what sort of infrastructure would we need
in order to make this technology safe and reliable and and then you'd start to look at the magic
roundabout in swindon and say hmm that's gonna be that's always going to be difficult for self-driving
cars but if we replaced it by a system of traffic lights for example then um then it might become
more predictable now if you do that you're also going to piss off some other people right and as
with any big government decision about infrastructure and transport you're going to have some people that
like it and some people that hate it and you're going to have a battle between those between
those people um because ultimately roads are quite political and the other thing i show my
students quite often is a zebra crossing right and a zebra crossing is a particular form of
infrastructure that doesn't exist everywhere in the world and is interpreted differently in those
places that it does exist but is a particular form of conversation that you have between different
sorts of road users right and you see it played out all the time when a driver approaches a zebra
crossing or a pedestrian approaches a zebra crossing and they have a form of negotiation
based on some what they would hope shared understandings about who has right of way and what and if
it works well people pass safely and the traffic isn't delayed for too long and everybody goes about
their lives right but that conversation is really hard for a self-driving car to do because it doesn't
do conversation because there's a weird connection between a driver and the person is out to cross
exactly the zebra crossing it's hard to put a finger up yeah and often the communication is
you know through a raised eyebrow or it might be a hand gesture all these sorts of things
now some self-driving car companies have tried to come up with other ways to deal with that
interaction which is where things could get really interesting so they've said well maybe what we
need is self-driving cars that communicate their intentions to a pedestrian and make them sort of
more more conversational or you say well actually zebra crossings are really annoying if you're a
self-driving car so can we just switch them all over to you know pelican crossings where there's
a traffic light and everybody knows when they're supposed to go that would obviously in some respects
potentially take more time than it would just to put the self-driving cars on the road especially
if they're up to speed somewhere like LA at the rate that AI is improving would would money just
prevail because this industry is going to be enormous to the first person that kind of wins
because it'll be their brand whether it be Waymo whether it be Zook that's training up on the streets
at the minutes let us drive around or whether it's you're launching their thing in the UK if you
were like a bus driver in London where if you look at a bus company it's got all the salaries
that it needs to pay but I don't mean London might be a Milton Keynes or a Sage Coast or a
Price in Swindon would you be thinking if that's going to give you a longevity of career right
now so I think buses are a really interesting one because again I think we can get stuck in the car
thing right and because most people most of the time their their their the go-to form of
transport will be a car and that will be the size of object that they identify with they maybe
don't think so much about buses but actually I think self-driving buses could be a really
interesting and beneficial form of the technology for all sorts of reasons right because you can
imagine them having dedicated lanes and you can imagine you know they do predictable things they
do the same thing over and over and over again you want them to be really safe etc etc um
it may well be that some bus routes are really good targets for automation right does that mean
that all bus drivers are going to be put out of work well probably not because at the moment we
expect our bus drivers to do quite a lot of different things and respond to all forms types
of uncertainty plus two other things that aren't just driving you know they also have a role in
if you're like policing the the the technology so I think a lot of passengers would be unsure
about getting onto a bus that had no professionals on it because if something started kicking off
with a fellow passenger you'd think well what happens here you said you think the scaremongering
of everyone's going to lose their jobs AI is going to take over and by 2035 70% of all vehicles on
the road are going to be automated or driving autonomously do you think that's an incorrect
statement I think it is yeah I don't I don't think necessarily it's scaremongering right so
statements like that often come from people who want to encourage the excitement about the technology
and say it's all moving really really fast and it's really exciting and you need to get on this train
right because it is it is sighted when 300 000 people lose their jobs well so that's it and
they're often but often the people who are giving you the most hype are are the ones that are least
able to deal with those the possible threats that rapid automation might might have if AI
is extremely extremely clever far cleverer than you or I because of the rate it is learning
and especially super intelligent AI then why do we need someone to work on the self-driving car
when does it become just AI works on the self-driving car oh well you see I'm not sure we are I'm not
sure anybody's thinking about that yet for the simple reason that the the context in which this
technology will have to operate and operate safely is so unavoidably human and I think that's why it's
so interesting that if we are not just to develop the technology but also to govern the technology
and make sensible decisions about the technology you need to understand a road a place in all of its
human richness right so you want humans to be there contributing to making that making that work
and dealing with the trade-offs that will always happen when you're trying to you know change
who it is that's able to to to move about in because ultimately you know roads are are shared
spaces it's not like um you know something like uh the tube network or somewhere in London which is
actually an entirely closed off space where it's actually becomes rather an easy engineering
challenge and you can imagine you know deploying AI systems to make the tube more efficient that
becomes quite an easy problem compared to the road space where do you think the people that have lost
their jobs are going to be displaced to though if there's no need for them in the technologies the
technology is cleverer than us so what I mean this is a classic problem of
of economic progress I won't say technological progress because often the thing that these
sort of disruptions come not from particular technologies but from economic disruptions
that that might be you know caused by changing business models or or changing economic structures
or changing patterns of immigration there might be all sorts of of things and one of the things
that modern societies has to deal with is that um is the sometimes rapid changes in people's
people's livelihood so I spoke about Uber I mean the changing economy of a London taxi driver
um has been pretty pretty rapid right and it happened that the in the grand scheme of things
you know the transition of Britain's economy from a manufacturing economy to a service economy has
had profound implications for how we think about careers and how we think about education and and
all of those sorts of things so yes automation will be a part of that but it won't be the only
part of that if you could squeeze the brake pedal on the deployment of self-driving car
features in the UK right now would you squeeze it or press the accelerator I so I tend to reject
the analogy I think the interesting analogy is with is with the steering wheel right so at the
moment I look at the development of the technology and I say it seems to be going in that direction
is that the best thing for society well I'm not sure it is I can really see some of the potential
benefits why don't you think it so I I think there are enormous potential benefits to make our roads
safer more efficient to allow people that currently aren't able to drive to move about
right massive massive potential benefits but at the moment the technology is largely being
steered by a few companies who I don't think have the public interest in mind right I mean
you know they have um sometimes the interests of companies and the public interest might align but
the job of regulators to be would be to make sure that those things are better lined up so
I think there's all sorts of things that government needs to do to um force companies to be a bit
safer to know more about when these technologies fail so that we can learn about how to make them
safer overall um I think we can do things like ensuring that the technology is developed in
such a way as it benefits people who currently don't have access to many forms of transport
rather than goes to places like San Francisco Los Angeles where it's actually that's where all the
money is there's a difference between you think and they will because often every human being in
England thinks the government should do something whether they actually do will execute a very
misaligned especially at the minute to do you think the things that should be done will actually
be done or do you think we risk the opportunity of a lot of very premature systems being allowed on
the road that have grave impact that we've never seen before so that is a risk I mean I do I am
relatively optimistic for what we're doing here in the UK you know I think there is the opportunity
for the UK to come up with some really good rules that mean that we do actually get the benefits of
the technology without deploying some systems that I mean it's not just deploying systems that
aren't safe but it's also deploying systems that then people end up hating which might turn them off
the technology right which would be which would be really bad and then they just think oh I don't
want to go anywhere near that whereas actually if you can bring people with you and and get people
to to trust you then it's good in the long run and so I'm I am pretty optimistic that the UK's
taking taking a good approach here but ultimately you know the challenge that you've identified is
a democratic one and that's that's what it comes down to for me as a social scientist is that it
this is a democratic question it's not just about leaving the technology to its own devices
so as a social scientist you'll know that many cars and technology is there already if you're
able to now sell a new car and that new car has to have lane assist technology on it has to have
anti-breaking technology on it because it's deemed that it's more safe for that vehicle to have it
then genuinely I find this question so baffling it's unbelievable and I actually had it with the
head of Devon's road policing I was talking to him about it which is this mobile phone
your car if you're in a Tesla funnily enough like the Cybertruck can you pick it up and try and take
a picture of the steering wheel if it's doing autonomous driving it will go berserk like it
will go mad that you've picked up the phone but like if the technology is there to essentially
go mental provide a warning that you've done something and then potentially
pull the car off to the hard shoulder and park up if you persist to do it right why aren't we
executing that technology yet if like one of the biggest deaths on UK roads is through
I mean so I think it's you know there are there are some really interesting questions like this
like like you know automated speed limits is one of the challenges I like to discuss
you know because it would be technologically really easy I mean you know the technology
involved in monitoring a driver to check that they're paying attention in the car is actually
extremely complicated the technology that says you mustn't go above the speed limit right we've
actually had that technology for decades and decades and decades why don't we use it because
you know it would alarm a lot of people they would feel like their freedom is being threatened
and there are particular things about cars and people's attitudes to you know the car as a personal
space and and feeling regulated in a car that that mean that it's quite hard to deploy some of these
some of these technologies I think one of the challenges that self-driving car companies will
face which I don't think they've really contended with is the fact that car culture is quite
libertarian right even though oddly you know the irony is that when we're in a car we're probably
subject to more rules than in any other aspect of our of our life we feel free and in charge
and car companies have been selling us this ideal of freedom right as because that's what a car
offers right it's great affordance it's great benefit is that a car can enable you to become
a free moving individual in control of your own of your own destiny which means that if you say oh
and here's a technology that will keep you safe I mean you need to look back at the campaign both
of us too too young to remember the campaign to introduce seat belts right massive resistance
to the idea of seat belts and people refuse to put them on big public information campaign just
to get people to wear their seat belts and then they got mandated and then they got you know it
that is now illegal not to not to wear them and that's a technology just you know that stops people
from flying through their own windscreens what's one prediction you're willing to make about self
driving cars before 2035 that we won't have very many of them in London that would be that would
be my my prediction well no hang on because had you asked me I may have said like five years ago
I may have said the same thing about American cities and the one thing that I've been taken
by surprise by and this is why normally I resist prediction because all predictions are going to
be wrong I've been really taken by surprise and impressed by the speed with which Waymo have
rolled out their vehicles in those cities not just doing a few journeys but having now substantial
numbers of of cars I was fast asleep in one I went to sleep for 40 minutes well it drove me
from A to B and I don't know there's nothing else going on in the car I mean that's amazing and it
woke me up bonging when I got there so that is the ultimate the ultimate test of a technology that
works right if it's doing its job it should be so boring that you you know lose interest because
I imagine I mean so let me ask you about this when you first got in it there must have been some
sort of sense of novelty some surprise oh my god yeah the first sort of the first time I took
when I actually vlogged the whole thing right because I do car stuff I post it on social media
why wouldn't I it's a significant thing to to go in a first vehicle that has no driver in it and
you're going to put your trust in it your faith in it and watch what it does is a significant thing
so the first thing that happened was it pulled up on the wrong side of the road
which I thought was almost comedic because I'm about to trust the thing to to get in it
and now I've now got across the road okay how did it not know that I was on this side the side
that it came in on is that it turned around and parked on the other side yeah so across the road
I went to get in it there was a sense of strangeness about the fact I opened the boot
put my bag in it and was going to close the boot before opening the door to get in the car like
is it going to drive off with my bag because you're so unfamiliar to it yeah um and all these
things creeping about the experience that you're so unused to in normal kind of day-to-day life
when in reality you could put your bag in an uber or a taxi in London you have no idea you could
just shoot off it's easy this is a Louis Vuitton bag I couldn't be god knows what but you at least
make some assumptions about there being a driver who's so yeah that you that you trust which is
crazy because really that Waymo has no other intention or heart than really to take you to
for me to be but then I got in the vehicle I got in the back because I think that when you're on
your own you have these novelty sense of importance rather than getting in the front of a vehicle it's
like you're being chauffeur driven by something yeah so I got into the back you tap the screen away
you go um and within two and a half minutes I thought was the best thing in the world yeah
because I was just like oh my god it's doing it I knew it was gonna do it I knew it was gonna do
it because I'd seen it do it on video yeah there's millions of them buzzing around I mean genuinely
every street you turn down now I'm not even joking every street you turn down now you see one they are
everywhere and it's just a really strange a really strange experience for the first time being in one
but one that A wore off really quickly but B convinced me that that is the method of transport
that I would take in in LA did you find yourself sort of judging its driving style and because I mean
you know about cars so I'm imagining you were interested in well how is it going to deal with
this or how's it going to go too fast or too slow um some friends of mine lent me a Jaguar F type
and it wouldn't fit my mrs bag in the boot it was on the final day when we had to go back from
our hotel to where I keep my other studio and that meant that I needed to figure out a way of
getting me and my other half back to where we keep the van studio say goodbye to everybody
close up and all the rest of it then go to the airport but I couldn't get this other bag in
F type so I ordered a Waymo on my phone and I followed the Waymo in an F type
now I can't confirm I can't confirm or deny and it'd be really interesting for people in the
States that have taken them to comment below if this is correct or true or not but a journey
that should have taken 12 minutes took 26 minutes and I almost felt like the Waymo thought it was
being followed or tracked and was doing cut some kind of anti-arrow and was trying to lose
technically I genuinely did start to believe that because it would kind of go down and right on
lane then cut across to and then go left and I sort of fascinated I've never actually looked
into whether it it was doing that or if it was just trying to avoid traffic in LA blacks up
really quickly or whether it thought this this random vehicle has been tailing me since we left
like why is that I found that quite so one thing that these these companies do which is totally
understandable right is that they're trying to reduce risk so all these the the companies have
a set of priorities at any one time and they and Waymo one of the good things I trust about Waymo
is that they prioritize safety over other things which means that a lot of the time they will
take a longer route to avoid really complicated junctions for example and it does mean that
a quite a few people have complained that this is not the quickest way to the airport right because
they might they might rather do like a series of three right turns than one left turn across
traffic in American roads and things and which which makes sense right but but you have to sort
of try and work out if that's the truth from the outside the only time I've not felt safe in a Waymo
for a few minutes is funnily enough when there wasn't somebody in it that could have broken the
law right and the reason behind that statement is my love had a window down in a group of
very American gangster looking lads saw her pretty girl like that was their full interest was my
partner in the car on a road not too far away from skid row in downtown LA we put the window up
there immediately shouting behind us wolf whistling everything the car did not know that was going
on the car it's just going to follow its route at its normal pace where if it had been me driving
potentially hired something quick I like my cars like I could have been in a cyber truck could have
been in a v8 Chevrolet so I'd have absolutely nailed it I'd have been gone I wouldn't have
given a damn about the speed limit I'd have just been gone then out of there and I think that kind
of ability to kind of break that curfew yeah when it's safe to do so in your opinion or it's going
to have a benefit to you it was the only time that I've not felt safe in one but also you feel
powerless in that moment right because you are in the hands of a system that you don't understand
and you don't know what the what the expectations were but the attractiveness is I'm in a very
fortunate position I travel a lot got some good businesses if I really wanted to choose an uber
driver every time to get around LA even if it was 40 minutes and 50 60 dollars I'm in a fortunate
position I could do so however even me feeling in that position being in that echelon couldn't
help in the end but add up the difference between what a waymo would cost right and an uber would
cost and worked out that over the course of two weeks in LA you're saving nearly a thousand
dollars on a lot of journey by taking waymo's because they're half the price compared to ubers
so that's the thing I think we don't we don't yet know about another thing that we don't know about
when it comes to self-driving vehicles is what it would take for them to be a cheaper option
because a lot of people will be seeing this as a nakedly economic calculation like they did with
uber you know am I going to take a black cab in London and we're going to take an uber well the
uber's half price right and then we also know that ubers over time got more expensive because
they reduced their subsidies because one of these the things that a new company will do
is subsidize these services while they're being tested or rolled out to see what the market will
take into attract customers and all the rest of it we still don't know how expensive a fully
functioning self-driving car system will be I mean what we know at the moment
Robotaxi $26,000 so but in terms of the the so you've got the vehicle right now a waymo vehicle
is a lot more expensive than a Tesla Robotaxi because it is absolutely covered you will see in
some really really expensive high-end sensors amazing sensors far better than anything on a Tesla
Robotaxi which means that the things themselves cost a lot of money now that maybe isn't a problem
if you're using them all the time and generating so much economic value day and night plus your
saving money from not having a driver but one of the things that we don't know and the companies
don't really want to tell us very much about is how many other people are involved you know what
the labour costs are behind the behind the scenes how big's the time so when you're in your waymo
one of the the other things that makes it trustworthy is that if it goes wrong
there are actual people available to remotely operate it right you've got people on customer
support you've got people who are able to redirect the car out of trouble we had an incident the other
day in Arizona where waymos were driving into floods right and they're remote operators because
the systems aren't good at detecting floods and you don't get very many floods in Arizona
so the remote operators had to redirect them out of out of there right so there are people
behind the scenes we don't know how expensive the whole system is the challenge if you're one of
those companies will be can you make the system more efficient so that you can guarantee that it
will carry on being cheaper than a than a conventional uber and we just don't know at the moment so
what is what do you think the biggest risk with self-driving cars is right now to humanity do you
think it's that terror impact how well we secure getting in and out of it but the system itself
i think i mean the thing i've observed the last 10 years or so as we've seen sort of society's
experiment self-driving cars is that the things that are gonna bug us are not likely to be
big you know high-profile crashes or massive terror attacks they're likely to be a series of
rather small boring things like the price of a thing like you know where can this thing go you
know like um will this thing park in uh there have been some incidents where in san francisco
they parked in the wrong place and pissed off the local decision makers and you just sort of think
it might be that the bugs people right it might be cars bricking in the middle of traffic light
junctions and causing traffic jams and that's not dangerous but it might really test the patients of
local residents and it might mean that the experiment with self-driving cars isn't allowed to grow
properly that's why i would like to see this better regulated so that we can get some of the benefits
but not necessarily drive into those uh dead ends what do you think the biggest myth is that you'd
like to kill as we close upstairs about it what what my perception or someone's perception of a
tesla model 3 walking running down the m4 and someone's not got their hands on the steering wheel
what do you think's like and i'm and i'm fine with that in the car next to it what do you think the
biggest like myth is that you like to maybe bust um tesla has its own set of myths but i don't think
that that's the you know because tesla is selling something called full self-driving technology that
is not full self-driving right but i don't think that's the um the biggest thing that i think the
most important thing to know about the technology is that a self-driving car is never going to drive
like a human and therefore comparing it to a human might not be the best way to to to make
decisions about these actually if we want the technology to succeed we're going to have to
recognize that robots will drive really differently and we'll have to make decisions based upon that
Jack i think there was absolutely fabulous you've given us an insight into something that many people
think about but only for a few minutes without knowing actually everything behind it around it
and what it could become so i want to thank you so much for giving me over an hour of your time
in the back of my van studio in the middle of london where potentially in the future 10 20 30
years from now but we never know based on la maybe sooner we'll be pulling up to do another podcast but
tobey will be in the middle seat rather than the driver's seat so thank you very much and i'll see
you then ben thank you it's been a pleasure
About this episode
Dr. Jack Stilgoe, a technology governance expert, discusses the rapid advancement of self-driving cars and their potential impact on driving jobs in the UK. He argues that while automation is progressing quickly, the complete replacement of human drivers by 2035 is unrealistic due to the complexities of urban environments. The conversation delves into the safety concerns surrounding Tesla's autopilot features, the ethical implications of AI decision-making, and the regulatory challenges that must be addressed to ensure public safety. Stilgoe emphasizes the need for responsible technology governance to navigate these changes effectively.
Original notes
Check out Tweak: https://www.tweakuk.com/
Join us for a deep dive with Dr. Jack Stilgoe, Professor of Science and Technology Studies at University College London, as we discuss the future of autonomous vehicles, the legal and ethical dilemmas of AI on our roads, and why society may not be ready for the self-driving revolution.
In this episode, Dr. Stilgoe—who advises the British government on tech policy—breaks down the difference between companies like Waymo and Tesla, the dangers of "Full Self-Driving" mode, and the complex challenge of assigning responsibility when an AI makes a mistake.
Don’t forget to subscribe to our channel for more exciting content about your favourite shows and celebrities. Hit the bell icon to stay updated on all our latest episodes
👍 Like, Comment, and Share this episode. Join our discussion in the comments section
Check out Tweak: https://www.tweakuk.com/
🔗 Follow Us:
Instagram: @Roadtosuccessofficialpodcast
@benedictfowler
Contact: [email protected]
Road To Success with Benedict Fowler
Tweak Marketing
0:00
84:37