Welcome to the United Nations | Department of Economic and Social Affairs

The following are the outputs of the real-time captioning taken during the Tenth Annual Meeting of the Internet Governance Forum (IGF) in João Pessoa, Brazil, from 10 to 13 November 2015. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the event, but should not be treated as an authoritative record. 




>> MODERATOR are we connected?  How is it?  Ah, that sounds good.  Great.  Good morning to everyone.  Very happy to be here early this morning with those of you who are in our audience and have been able to make it to join us for this very interesting and innovative presentation and discussion.  I hope we can all move up as best we can, because we are so few, that it would be great to cluster and actually have more chance to see each other as we talk.  We have a few more folks coming in.  If you can move up front, we'd love to have you up front so we can take advantage of all the folks that are here.  Good morning.

So, welcome to our session today, the IGF, which is a session that builds on work that was done in Istanbul just a year ago by several of OECD's expert staff, notably Rudolph Vanderberg.  My name is Tracey Weisler.  I'm coming in from Washington, D.C. and I'm a member of the U.S. Federal Communications Commission and I have to tell you that it is an irony that I am chairing this panel or moderating this panel because I grew up in New York, which means I never touched a car until I was 21.  However, I also went to graduate school at the University of Michigan, and for those that do not know Michigan, it is the home of Ford and General Motors, big car companies that enjoy quite a big presence there.  So I came to understand later in my life what importance the car has not only to the American person in their life and their mobilities, but in their psyche, in their actual spirit.  So it's an interesting conversation for me to moderate today.

I think this is an innovative panel discussion because as it builds on the IoT work that was featured last year, it also narrows the subject matter to something that's actually a very provocative in and of itself.  So while we can say things like:  By the year 2019 we'll have 10.5 billion devices, it's estimated, connected on the IoT space, you also say when you look underneath the categories of the IoT and look at different subsectors, different submarkets, if you will it becomes even more of a provocative conversation the issues get even more deeper and more detailed and more treatable, hopefully, by all of us here in this space.

So what I'd like to do is tell you what we'll do in the next hour or so together.  We'll start with Bill Woodcock who is to my right.  He is CEO of ecotruck which is a vehicle manufacturer in California.  He has been doing a lot of work on the Internet side prior to this, IVP6 transition and other kinds of important work.  But he will take us through a scene setter.  We will set the scene of the discussion, looking at autonomous vehicles.  You break it down into the classes.  And as you look at the issues that might be teed up by their emergence, I guess, in the markets.

Then we'd turn it over to Alexander Lu who is here on my left.  He's the strategic Director of BYD Brazil.  This is the largest manufacturer of rechargeable batteries.  It is an expert ‑‑ he's an expert in bridging companies between China and Brazil.  In fact for the past 20 years, you worked in both of those countries quite diligently.

Then we are going to turn it over to a remote participant who's actually, I hope, hooked up adequately to be able to participate when the time comes.  He is Guillerme Correa, I think that's based in Brazil, yeah?  He will be phoning in, Skypeing in, hopefully, to chat with us about his view of the public policy side as he is an infrastructure analyst at the Department of Science, technology, industry.  He's also manager of the M2 M IOC centre here in the Brazilian ministry of communications, so he has a policy perspective that will be very useful to us.

And then just to finish us out, we'll have Helani Galpaya who is here to my far right who is going to be taking a different look at some of the issues that will be presented earlier.  She is CEO of LIRNEasia, a pro poor market think bank working across the Pacific on regulatory issues of the so her recent research and policy has been related to net neutrality, e‑government so I think her perspectives will be very useful today.

So with that, I'd like to start out with Bill, and he has a few slides he wants to show us to get the conversation started.  Thank you, Bill.

>> BILL WOODCOCK:  Thanks.  So I'm just going to try to run very quickly through sort of an overview of the issues so that you can get a sense of what we're going to be talking about, what kinds of things may be interesting to talk about in Q & A and discussion.  So our slides?  Good, yes.  Okay.

So first of all we have, to lay out what's out there right now.  For autonomous and connected vehicles, basically we've got cars and fleet trucks running around on the ground, and we've got fixed wing aircraft of stuff that's not really common yet.  There are sort of boats, although there is a lot of interest of doing automated container ships, right?  Container ships without crews.  Submersibles because there isn't a lot of market for submarines in the first place.  Rote oh craft, helicopters because the control systems for helicopters are a lot more complicated than fixed wing aircraft, and the danger tends to be more.  Although the qualcopter are obviously the more common form of UAV now.  And then lighter than aircraft.  They're like submarines, there isn't a whole market for wimps and derges(?)and stuff.

So moving ahead, so the concerns that people have, the kind of reasonable ones are privacy, because now we've suddenly got a whole lot of moving camera platforms moving around and lots of sensors on them reporting data back.  Safety decisionmaking.  Right now we're really used to people making constant decisions about:  Do I need to step over that thing in order to not trip over it?  Do I need to steer this way in order not to run into somebody?  Is there likelihood of somebody stepping out from behind that parked car?  All those things are running through your head all the time.

But with vehicles, all of that has to be programmed by somebody.  And the person programming isn't in the situation that they're programming about.  They're trying to imagine it.  And you're never going to be able to completely imagine it.  So there's a big issue there.

Airborne UAV failures.  So a flying autonomous vehicle, when it has a failure, there's nobody sitting in there on board to make a decision about where a safe place to land is.  If we suddenly have a lot of these instead of just a few?  If they are operating out of line of sight of a controller, these become bigger safety issues.  Things falling out of the sky onto people's heads.  If there's enough of them, it becomes a safety issue.  Code complexity is always an issue with everything that depends on software.  The more complicated the situation, the more complicated the Code to control it, the more complicated the Code, the more likely there are to be failures and problems in the Code that have to be addressed over time that are found through problematic accidents.  And software vulnerability.  Anything big has a bigger tax surface if it's not designed really carefully.  If hackers can get in your car and cause it to drive around, this is possibly more entertaining for them than getting into your computer and causing it to sit there and beep or something.

Then there are kind of outlandish concerns, the things that people who don't really know about the space worry about because they're not thinking quantitatively about it or they don't have the background.  So terrorism is the number 1 thing there.  This is something that I hear a lot about from U.S. government policymakers is:  If there are UAVs out there, can they be used to do terrorism?  And fundamentally UAVs are already very common in two spaces, military, where they're used to blow stuff up and civilian where they are used by hobbyists to have fun.  And so all of the new activity is in the commercial space, companies that want to use these to do business.  So companies doing business are not primarily your terrorists that want to blow stuff up.  Militaries do, maybe individuals do.  Those are the ones who already have this stuff.  So probably terrorism is not going any new issue here.  Availability of air space.  Drones driving around.  Are we going to run out of air space?  Are they going to collide with each other.  There's lots of air space.  3 dimensional space is more confined than 2 dimensional.  People think traffic jams there won't be in the air.  AI mortality.  Are machines going to get really smart and start making decisions for us and about us and so forth?  No.  The machines are not smart.  They just execute software.  The software is written by people.  The decision rests with the person, not with the machine.

So one of the big things that is not necessarily common knowledge to people who buy cars, for instance, but is fundamental in the auto manufacturing industry is this distinction between safety and entertainment.  So you think of your car as having a computer in it.  And you press the touchscreen to see how much gas you have left to change the station on the radio to, to play an MP3.  To check the tire pressure, all these different things come at you through this one screen.  But in reality, we've got two completely different regimes of software and computation happening inside the car.  On the safety side, we've got things like lane detection.  Are you middle in the lane?  Drafting?  Are you following another vehicle very closely, too closely for a human would be able to control breaking?  Collision avoidance, is something coming at you that needs to be avoided?  Hazard detection, is there something out there that you need to steer around?  Accurate navigation?  Is there a bridge over the river on this street, or is there not?  Driver sleep detection, is the person in the driver seat starting to fall asleep?  Is their attention wandering?  There are vehicles now that automatically sense that and will try to wake the driver up if the driver starts to fall asleep.  Autonomous driving.  This is when the driver's reading a newspaper, the car is driving itself.  Tiny things, automatic wipers if it starts to get misty.  Will the wipers start automatically?  Telemetry, is there data being collected inside the vehicle, maybe about the engine, maybe about the environment that is being relayed back to the manufacturer or somebody else?  And error reporting, if something goes wrong in the vehicle, if something breaks, if you're out of gas, if you're out of window washing fluid, if the brake pads are getting thin, is that information being conveyed back?

On the other side you've got all the entertainment stuff.  The number 1 entertainment in vehicles is talking on the cell phones, right?  So vehicles for both safety an reasons and for convenience reasons and for keeping the customers happy, they really try to make talking on the cell phone easy.  Audio, listening to audio, music, whatever.  Video.  There are plenty of places, particularly in Asia where you look in the a traffic jam and half the people are watching soap operas.  Video in cars is going to become more and more prevalent as people want to be entertain.  Likewise entertaining, world of war craft while they're driving?  We got to figure out how to make that safe.

WiFi hot spots.  My wife's car, you get in, there's a WiFi hot spot for your otherwise unconnected devices.

Per driver presets.  When one driver gets in and sits down, does everything in the vehicle change around to accommodate what that driver is set, and somebody else gets in that shifts around again.  Dashboard visualization.  Are there pretty animated pictures coming up on the dashboard to help you understand the different aspects of your vehicle?  And then active audio environment.  One of the sort of more controversial features now of performance automobiles is ‑‑ is this not working?  Is this one?  Oh, closer?  Sorry.  Active audio.  So one of the more controversial features of performance vehicles now is much of the engine noise of the vehicle is being generated by the entertainment system inside the car to make the driver feel like they're driving a performance vehicle.  And many of the manufacturers are getting kind of caught by reviewers, pull the plug on the stereo system and suddenly their big muscle car is no longer making any noise.  Conversely, luxury cars do audio sensing and generate inverse waveform to do audio dampening to make the inside quieter.  So this stuff is also in entertainment.  It doesn't make the person inside safer.

So these two regimes are treated very differently.  Safety stuff tends to run on realtime operating system.  The entertainment stuff tends to run on some other variant, whatever is at the moment to start the project.

The safety stuff, they roll out one feature at a time really carefully and they go over it with their lawyers and they look for test cases and they try to establish caselaw.

The entertainment stuff, everything in the car might change between one year and the next year.  So big example of that is Ford got rid of Microsoft as the supplier for all of their dashboard stuff, their entertainment stuff as of the current model year.  For the four or five years before that, every Ford with a touchscreen was run by Microsoft software and it was causing a lot of complaints.  And so the complaints started coming in as soon as they started selling cars.  So at that point Ford started a parallel effort to replace everything Microsoft had done and it took them about four years, three years to get there.  But they had no hesitation to just swap out the entire thing because there's no liability associated with that as opposed to all the safety stuff.

So let's talk about occupied vehicles, versus unoccupied vehicles.  Occupied one that contained a human or something unique.  Unoccupied vehicles are something that don't contain humans and the stuff inside is not really worth more than the vehicle itself.  So why are these different?  An occupied vehicle has to make "moral" judgments about the relative value of human lives, the lives of the people inside, the lives of the people outside, the lives of the people in other vehicles coming towards you.  And these are decisions that have to be encoded in software.

Unoccupied vehicles, on the other hand, have to make decisions about the tradeoff between their own property value and the safety and security of the people around them, right?  So if somebody runs out in front of the vehicle, does it crash itself into another vehicle, destroying two vehicles, one of them belonging to somebody unknown, in order to protect a life of a person who might or might not have gotten hit, right?  At what point does that tradeoff happen?

A big one.  Connected vehicles right now come from the factory with a sim just like a soft phone does.  The sim has an account associated with it with a particular carrier.  And the bill is being paid by the manufacturer.  So there's a certain price below which a connected vehicle is not going to go because in a $500 scooter, the manufacturer doesn't want to be on the hook for another $200, $500 a year for the data plan for that scooter, right?  For as long as a person will keep the scooter.  But on $100,000 luxury car, yeah, no problem.  So the problem here is that once that's out the door for the manufacturer to change data suppliers, to change carriers, they have to do a vehicle recall.  A vehicle recall, everybody bringing their cars back in to dealers' shops is incredibly expensive and horrible PR.  So manufacturers try very hard to avoid recalls which means effectively they're completely locked into the carrier that they were using at the time the vehicle left the factory.  As long as they can contain that problem and make it a problem that only affects a small number of accounts and vehicles and costs, it's know the that big a deal.  But for connected vehicles to reach a lot of people?  This issue has to be addressed.

Smartphone manufacturers that have a relationship directly with the customer, like apple and to some degree Samsung are in exactly the same position.  They want to maintain the relationship with their customer and let the customer switch carriers.  And so it's very likely that we'll wind up with a sort of reliance of high‑end smartphone manufacturers and automobile manufacturers to drive soft sim regulation to make soft sims legal.  Soft sims are where you can just change the software who the carrier is rather than physically changing a sim card.  The sim card is not adding any value it's merely a way for them to lock in the customer and make it difficult for the carrier to switch carriers.  So moving from cars to aircraft, UAVs are becoming cheaper, more reliable, longer range, able to carry more stuff, so there are going to be a lot more of them.  The current regulatory regimes were all created where there were very few remote control model planes run by hobbyists.  That's not the situation we're moving into.  And military, which is always exempt from regulatory regime.  So we're in this situation globally ‑‑ interesting situation globally right now.  If you look at WiFi, there is this ban of allocation in the U.S.  The U.S. was a big market.  So WiFi chip manufacturers made chips that would talk on frequencies that were available for use in the U.S.  But they made a lot of chips.  And those got put into a lot of devices.  And the devices, because they were the cheapest devices they could mass produce, got sold globally.  So you had devices speaking on U.S.‑allocated frequencies in South Africa where the same frequency might have been allocated by their regulator to garage door openers or to the military or something else.

And so now you had interference from consumer device that was very broadly used, and they had to change their regulatory structure in order to accommodate something that happened de facto.  It happened through the market.  Governments aren't used to having to do that.

So at some point very soon, some government is going to say:  Here are our regulations for UAVs.  And they're kind of complicated.  And they can be implemented in software or in firmware in ship and a bunch of Chinese factories will say okay, we understand this and it's simple enough that we can do it and we're going to put it in a 2‑cent chip and we're going to sell a billion of these chips.  And those will all go into devices.  So UAVs, you know those selfie sticks, selfie sticks are going to seem like an anachronism very, very soon.  So instead of a selfie stick, you will take a camera, throw it up in the air and it will follow you along and position it for the best shot of whatever you want to do.  And these are already happening.  Those devices are already coming on the market.

When your camera is the thing that flies around near you, they're going to be a lot of these.  And people are going to travel and take them with them.  These are not going to obey different regulatory regimes in different countries, which means some country is going to be first and their regulatory regime is going to have to be adopted by the entire rest of the world whether they like it or not.  Same as WiFi.

Okay.  So I'm nearly done here.  What kinds of data gets communicated?  Right now, we've got vehicle location information.  We want to know where you park your car but you can't remember?  There's an app on my phone that tells me where I parked my car.  Telemetry and error reporting.  So if something goes wrong on a car, the manufacturer finds out, I find out.

I'm set up because my parents are kind of elderly, I'm set up with an app for their car, also.  So their tire pressure gets low, my phone tells me and I could take care of their car to the station and fill the tires there.

Remote controls.  So you're going to hear a bit more about that from Alexander.  But obviously with UAVs, there are times when people want to control manually and they use the remote to do that.

Traffic speeds.  So there are feedback loops in mapping systems.  So a lot of mapping systems now have not just the mapping but how congested is the roadway?  One car drives down the road, car behind it needs to be able to benefit from the sensor data coming in from the car in front.  There are a lot of situations where that's going to become more prevalent.  Traffic data is the very first and simplest one.

Map improvement.  Most recent tesla upgrade includes ‑‑ every time tesla drives down the road, instead of the road being represented as a path, now every lane on the road is being mapped individually and the sender of each lane is being mapped as a probability function.  So the assumption is that tesla drivers mostly drive their car kind of near the centre of what people think of as the lane, right?  If beam drive down the road, every single lane will get mapped.  People find it easy to change lanes, that all will become part of the data.  Places where nobody ever changes lanes for whatever reason, the car probably will not change lanes for you there, either, because people didn't do that.

Toll tags.  Toll tags are one of the very first and simplest and easiest connected vehicle things.  Just slap the little device in your car and it responds to radio stuff and debits your account when you drive across a bridge or through a congested area or whatever.

Also, they are used for traffic monitoring, right?  So that the people operating the system know what to sell when, what to sell to people or governments or other governments, marketing purposes.

Very large there's very little security around toll tags.  You don't need to be the authorized party to read the toll tag of a vehicle that drives by, by and large, which means interactive billboards could be watching toll tags and coordinating vehicle ownership with somebody who bought something last year.  The ads you see on the World Wide Web that are tailored to your interests?  Yeah, well billboards may be like that soon.  And nothing beyond what they're around is necessary to do that.

Garage door openers.  That's another obvious early connected vehicle state, right?  The little thing that you press to get the garage door to open, those are built into the visors of a lot of cars now with using a standard, an open standard.  And entertainment.  If you want to stream data.  Sirius XM satellite radar.  Or tuning into a television show or whatever.  In the future, we will have vehicle‑to‑vehicle communication.  Stuff that's not coming from out on the Internet but coming from the car in front of you or behind you or next to you.  Sensor data.  Data about what that vehicle is about to do so you can coordinate your activity with that.

In the UAV space, flight plan filing.  This is the big regulatory thing.  Will commercial UAVs that fly out of range of a human controller be required to electronically file a flight plan as they take off?  So the vehicle starts to take off, it electronically says where it's going.  And the aviation authority would have an opportunity to reroute and tell it, okay, but you're going to need to avoid these other things.  There's something happening here, an plane is landing at an airport, television new crew is taking off here, take this path.

Interactive congestion avoidance.  Your car would reroute itself in order to avoid traffic and avoid creating traffic where it's avoidable.

Things to watch out for, personally identifiable information.  Stuff about you that's being shipped by your vehicle to parties unknown.  This is probably not a good idea.  Information relating to content or packages of vehicles.  Let's say you cross a national border and there's something in the truck of your car that was legal in the first country and it's not legal in the second country, do you really want your car ratting you out?

Likewise somebody gets into your car and, do you really want a billboard to change based upon who your passengers are?  Do you want that to be part of public record?  Who sat in your car?  And then sort of unnecessary oversharing of any correlatable sensor data, right?  Your car drives past somebody who is doing something on the sidewalk and is filming continuous video that's improving maps but also happens to capture somebody doing something that they thought they were doing in private, right?  These are all issues that we kind of want to avoid.  So what governance changes?  The big liability for the manufacturer or owner.  Manufacturers don't like this new source of liability.  And they are desperately trying to find ways of offloading that liability back onto the owners of the vehicles.

What behaviors will be needed to required to protect human life?  And what will be required for convenience of people?  Like is the convenience of a robot car more important than the convenience of a human?  Most people would say robot cars should get out of your way if you're trying to get somewhere.  It shouldn't take the parking space you want.  How far is that kind of thing going to go?  And then for connected passenger vehicles, the big question is:  Dream driver entertainment versus how much can the car do for you?  Right now it's illegal to talk on the cell phone unless you have a headset, or it's not, other places are not illegal.  Or it's not illegal if you're in the moving car.  Wide variety of regulation in that area right now.  In the future, it will be legal to red a newspaper while driving.  But the question is:  How tied will that be to the capabilities of the vehicle you're in?  Obviously it doesn't make sense to read a newspaper while riding in a cars that are by and large in the market right now.  That transition is going to be gradual.  Not every car is going to become capable of smart driving instantaneously.  So the law is going to have to take into account that different vehicles have different capability, different degrees of autonomous safety.  So depending what vehicle you're in, what you're allowed to entertain yourself will be different.  And that's a tricky thing to regulate because it has so many different permutations.

And then one of the issues is passenger aircraft, commercial aircraft.  What are you allude to do inside the plane?  Talking on your cell phone?  Having your cell phone on during takeoff and landing using an iPad versus using a laptop.  Again we've got this wind range of regulation depending upon the airline, depending upon what country you're taking off from or landing in.  Some of it is safety oriented.  We don't want RF interference with control signals and wires in the plane.  Some of it is purely about the sort of comfort or convenience of other passengers.  Do you really want the guy sitting next to you yakking on his phone for 12 hours while you're trying to sleep?  Right?  Some of this is important for governance and regulation, some of it is going to be just rules.  And the safety side of this is all about the RF and the surfaces and software.

Last slide.  Infrastructure.  Infrastructural changes.  One of the things that we were talking about a little bit before we sat down here is that a famous science fiction author once said that the future arrives but it's distributed but not evenly.  So a lot of this stuff is going to happen in Germany very soon, right?  We've got the infrastructure changes, the regulatory changes and the vehicle changes.  And they're all coming together pretty quickly.

Japan, same thing.  There are other places that are going to take a very long time to have any of those three things.  And they kind of, depending upon how much and which of those three things happen, different degrees of safety and convenience are going to be achieved.

For cars and trucks, machine readable identification, parking and traffic law signal.  The vehicle needs to know where it is and what's legal there.

For charging?  There's a very high correlation between electric vehicles and smart vehicles, and that will continue more and more in the future.  You want vehicles that start to run on power to be able to charge themselves.  That means conductive charging, so they don't actually have to mechanically couple with something or sort of smart mechanical coupler to do the charging.  That infrastructure basically doesn't exist except for a very few pilots and trials right now.

For UAVs, safe landing spots.  If you're in an apartment building and you want to receive a delivery via UAV, then Amazon needs to know what balcony do they land on to drop your box off?

What happens if a UAV starts to run low on power, it doesn't have enough power to get back to its origin and recharge and do whatever else it needs to do?  It just needs to sit down for a few minutes and park until it's ready to do its next thing.  Loitering.  Where will it be legal for UAVs to loiter?  Same thing with cars.  You don't need your car when you're in a meeting?  Where does your cargo and park itself?  What spots will be allocated for use of robots?  How difficult is it going to be for the devices to get to and from those?  How much will human convenience be traded off?  And then emergency landing.  Will there be designated emergency landing spots for UAVs?

All of the protocol interactive flight plan filing stuff, none of that exists right now.  It's just talk.  It's not real code.  And the sort of collision avoidance cal classes stuff that ‑‑ calculation stuff that aviation authorities are going to have to do, again that's all just talk.

So sorry I took quit a while.  But I was trying to give you an overview of the issues.  Back to you.

>> TRACIE WEISLER:  Thank you, Bill Woodcock.  I think Bill did a great job setting the stage here because the landscape is very broad and as I'm listening to it, I react as both a citizen driver and also as a regulator and I think what he did was show how each particular decision tree leads to yet another decision tree on issues regarding legal liability, security and risk city planning, use of these potential automated vehicles for the disabled community and what that means on a daily basis.

So my questions are starting to spin.  I wonder if the panelists have anything they want to add right now at this time?  Any thoughts right away?  Any questions for Bill?  Okay.

Well we'll move on to Alexander who is going to give us an overview of the operationsal side in his world, the BYD.  Thank you.

>> Alexander:  Hi.  Thank you, Bill.  Your presentation covered about everything about the ‑‑ I will explain shortly about the BYD and how ‑‑ what we are doing and what we have now and what we are planning for our new future.

BYD is a Chinese company that don't have the government and shareholders but have full support of the Chinese government.  We are very central to Bill's presentation because unique companies the automotive car but before that we are IT manufacturers.  So we produce cell phone tablets and notebooks and everything before and now we are making cars.  So this makes our team very open mind today try all kind of technologies.  We tried entertainment, awareness, we also have a lot of projects for autonomous car.  BYD was considered this year one of the we have two things that the car.  Our main business is battery manufacturers.  And so we purchased in 2000 car and changing all for electric and hybrids.  This is our institute for research and development.  And some of our entrants.all of the projects you are seeing, all of the car/buses we already have working for several years, and now all the buses are electric.

For the cars, every car, fully electric and the rest are hybrid and electric vehicles.  It's very interesting because for the autonomous, all our cars have the ‑‑ is connected to grid.  And when you have some problem as they can download from, they save the data and we can receive it when all the information for any bugs.  And so they send to our centre and we send the car to fix for the client.

This is auto lines of products we are putting in electric motors.  We have like buses, trash trucks, cement and even mining companies that we are putting all electric vehicles.

And for the passenger cars, we have now the control remote.  Some stuff that we already have.  This is from ‑‑ that is cars that ‑‑ Su Rui that have cars that ‑‑ lower the volume, please.  And this car is our last version car that is for gasoline.  And this already has remote control.  We've been putting this as optional for all our models.  And you can drive.  Driving range is like 10 meters, 15 meters.

I understand this remote control, you have a lot of sensors that can give like security to avoid collisions or for the guy with remote control tool, not to hit somebody.  You can find a lot of videos from Internet from people who have the remote control.  Say like parking by themselves.  They like driving.  And it is very cool.

Okay.  I'll pass it.  Thank you.

For the autonomous vehicle, BYD make joint venture with Astar Singapore.  We basically have several cars running in telemetrics all over the city.  And they are collecting the data to create artificial intelligence.  So we are using for our passenger cars.

Well, Bill has already told advantage of autonomous cars.  This is like for everyone can use.  Children, everyone that can go in places and just sit in car and take you wherever you want.

What he said also is when get to school, the car can go any place else to carry others, other people.  This is like a revolution with public transportation.

BYD concerned about security, hackers.  So basically we put all the artificial intelligent used all by the internet.  That's why we are maintaining the test in Singapore.  To avoid hackers driving the system and drive the car away.  So Internet parts we are using main for containment.  And we have lack of regulation for this kind of autonomous cars.

Right now, many countries it is not allowed to drive, to have a car without drivers.  And others issues we don't move forward with this car that we already have that is almost ready for the market is because that we don't have legal, we have a lot of legal issues for the AR decision.  We are afraid that some other cars hit them, instead of suing the driver, they sue the manufacturer because AR was driving.

Basically I have this video that help.  This is lots of tests we running.  Cars that are autonomous cars that are driving.  You can see a lot of people have bicycle.  So traffic in China is quite crazy.  And this is what we have autonomous car.

Now BYD is a conservative manufacturing company.  But we start to branch a few years.  And soon we will launch a lot of new cars in the market.  Okay.  Sorry.  That's it.

>> TRACIE WEISLER:  So, we have Bill Woodcock here who's laid out a pan openly, a huge range of issues, and I turn over here to Alexander and I'm seeing that I'm ready to buy one.  So this is a very dynamic because my intellect is aware of the issues that have been mapped out greatly on the right, and on the left, I have a wonderful advertisement for the future.  And so this is very interesting.  I don't know what you're thinking, but I'm going to now ask that we try to get our remote participant coming in, Guilerme Correa, I hope he's ready to chime in because he wears his government hat.  I would ask him as a government official, how would you reconcile this world that's here today and the world of future issues and see what he has to do about that.  So let's see if we can patch him in the he's ready for us.

>> GUILHERME CORREA:  Hi, are you hearing me?  First, I would like to that the opportunity to be here.  We are together the team of the ministry of communications we had some technical problems in the beginning of the problem.  We can't hear Bill's presentation.  But we think that things that will this happen the future should be thinking in the future because today if you have an accident today, the law is ready to judge who is the guilty?  If the car breaks, we can find who is the guilty today.  In the future, we will have sensors, actuators, softwares and probably we think that the law, we will find a way to find who is the guilty the same way we find the guilt today.  That's our opinion.

We are trying to work together, here in government we are trying to work together the companies, the initiatives and we are trying to listen to all the opinions to start a new regulatory and that's it.

>> TRACIE WEISLER:  Thank you.  Can you hear us?

>> GUILHERME CORREA:  Yes.  Well, I think it's clear that these policymakers sitting inside a government ministry, he has been charged with handling immense liability issues as well as manufacturing and personal side and so we'll get back to those.  I'm sure there are questions we'll have among ourselves with the presentations.  But what I would like to do is use Mr. Correa's concerns and also his call to work together as a multistakeholder approach to this evolving technology.  I'd like to turn to Helani Galpaya because I think she can bring us back to what this means on the privacy side given your work with both southeast Asia and ‑‑ what does this mean to you when you hear these conversations developing and the worry from the ministry about how to unpack some of these issues?  What are some of your thoughts on the matter?  And what are some of your concerns about the challenges?  Thank you.

>> HELANI GALPAYA:  So first reaction is a lot of this sounds, in a way, like a pipe dream and I'll clarify that.  I mean today, a lot of south Asia, some parts of southeast Asia are unable to handle liability in a simple road accident.  Really, really poorly handled.  Because actually the laws don't exist or they are not clear or they exist but there are informal systems that take over.  So that's the situation we're dealing with.

So a fully autonomous driving system, so not just cars, because everything else needs to come together as Bill pointed out, seems not very realistic for a lot of countries in the near future.  But there are possibly exceptions like Singapore, right?  Where I see maybe if they want, it could be one of the earlier adopters because there are systems that generally rule‑bound systems.  When you have rule‑based systems and somewhat closed systems, a lot of things can happen, looking at it from the a system point of view.

However, it's not like we do not have sims embedded in things in Asia.  Some of them are static.  So like your electricity meter, if they are prepaid, they will most likely have the sim inserted to send information back and forth.  Or your solar panels, which is a big concern for electricity, a lot of solar.  They will have chips.

So this will increase.  It is becoming quite common when you have a highway, tolls.  So that is there.

So the biggest sensor in the mobile phone, the mobile phone people walking around with a mobile phone and that data is being used for analytics.  We are doing as ministry, traffic mapping, city planning, all of that work, several information are being generated and large data analysis.

Around all an of this, the conversation that we are having about who owns the data, is my data, the Telco data, can that Telco now share it with an advertising firm as an example you gave to push ads to me?  And so on and so forth.  Is absent in a big way.  So you see it here.  You see it being written about in more developed economies and it is absent.

And partly that's part of the policy process and the people in power haven't caught up to it.  But interestingly in surveys when we ask people about privacy, how concerned are things like even your phone numbing given or your Telco or your government knowing about your browsing habits?  Do you know privacy is experiential.  And the lack of it, this is a real black and white thing that hits them as a violation.  It's not something that a lot of people think about.  And this is in representative surveys of poor people in multiple countries or even microentrepreneurs.  It's really interesting.  So that conversation absolutely need to take place.  And I suspect it will take place in the context of cell phone data in the context of other moving telemetry in Asia.

What information goes back and forth what kind of networks you need to have.  So when you have promised feeds of X but actually delivered through .5 X, optimistically .2, you get 20 percent of the speed that's promised, this is all GSM networks, mostly, the ability of the networks the handle the high bandwidth communication is problematic, reliably.  Of course it can handle it.  So I was talking to Bill.  And a lot of this stuff is probably thin in terms of data requirements going back and forth.  Then it's not the through put that you're actually getting, it's the actual availability of the network in two ways, which is a lot of times you actually are in an area where you're supposed to get signal but you don't get signal.  The network is gone for whatever reason.  Or you can't get on.  Or some parts of countries, even though we are getting very good about coverage, 95 percent of the conversation, it's incredibly patchy.  And road networks are extending everywhere.

So the difference between what Bill and I were talking is data that possibly can be stored until you have a network connection like Bill was saying, that's fine.  But realtime data to avoid accidents and stuff like that, I'm not sure how that's going to run on these networks, certainly not on the public Internet as it is right now delivered through mobile networks.  A private specialised channel and frequency, perhaps.  But then that brings out a whole policy issue about what frequencies are going to be and who gets to own it and stuff like that.

>> TRACIE WEISLER:  Okay.  So now we've had Helani talk a little bit about her skepticism and how it relates to the network side as well as her feeling or her potential sort of forecast of how this all work in the region of southeast Asia.  We still have Bill Woodcock here from the whole list of challenges.  And then we had BYD Alexander showing us the future.  And now with our government official coming back, I hope you have questions for the panel.  I really believe most of what we talked about resonates with you, you have a feeling about some of these issues.  Even if you are a driver, I know some of you are technology and other experts, but just from the driver's point of view, there are questions and concerns.  So what I'd like to do is I'm going to actually try it this way, which is take the mic to you.  Let's see if it works.  Are we on?  It's good?  Okay.

>> Thank you.  Very interesting set of presentations.  So as I'm listening to the stories of connected cars, I start thinking about let's start with parental controls.  So I put the child in the car say go straight to school.  Come back.  Park in the garage.  Fine.  I can imagine that works now with all kinds of other things.

Then I can imagine somebody driving down the road and the police or the boss or somebody decides that they really need to see that person now.  They go in.  Take over the car.  The car drives to the police station.  How do you deal with all the individual autonomy issues and the fact that somebody else can actually take over a vehicle?  It's one thing for commercial vehicles like trucks where the company needs to know where stuff is and directed to places.  It's very different for individual personal vehicles.  Thank you.

>> TRACIE WEISLER:  Alexander, do you want to take a try here?

>> ALEXANDER LU:  Sure.  First step is to make the car drive to school.  And second step, we have like integration with public transportation because they need to have this control also.  Talk more about it because once you get over the ‑‑ like, for example, the police needs to stop the car.  They should have control for that because ‑‑ don't understand signals, so we need devices to control over the cars that you have running on the streets.

>> I think DVD players are an area where that particular fight has already been fought and in some ways lost by the consumer.

>> BILL WOODCOCK:  The big question here is if you buy something, do you own it?  Do you have the right to control the use of the thing that you own?  Or does somebody else have a prior right, senior right over yours, even though the property right is yours; right?

So with a DVD player, you bought a DVD, you bought the DVD player.  You put the DVD in the DVD player and it tells you forwarding not permitted.  You can't skip over the ads that tell you what a horrible person you are if you have more than three people watching the DVD.

The same kind of things are happening with consumer goods.  There's a huge fight in the agricultural industry over tractors.  Tractors are the single biggest investment that most farmers ever make.  They cost many millions of dollars apiece now for big ones.  And the farmers, there's a big open source movement in the farming community to try to replace the operating systems of tractors and a lot of cryptography goes into preventing that.  These people are spending vasts amounts of money and being told that they cannot control the thing that they own.

So, yes, it's going to be a huge issue.  Who owns the vehicle?  Who gets to control it?  The fight between those two parties.  And it's not just a question of under ideal circumstances can the ‑‑ police commandeer your vehicle someone wanted for questioning?  That's one question.  Another question is can a hacker take control of two cars that are supposed to be off circling around waiting for their occupants and use them to herd a third car to make that third cargo somewhere where it doesn't want to go by threatening it with collisions if it doesn't comply?

There are a lot of problems that will have to be addressed at some point, particularly in an environment where all of the vehicles are this way.

So in the video you saw where there is a smart vehicle that is navigating crazy, complex obstacles and people on bicycles and people looping out from behind things, and people talking on phones as they cross the road, that's one environment.  And it has its own set of dangers.

The environment where all the other vehicles are also under autonomous control, have lightning fast reflexes, this is the automated that has a different set of dangers.

>> TRACIE WEISLER:  Helani, do you want to join?

>> HELANI GALPAYA:  You're giving nuances of her example.  I like that.  But the first most obvious example, I mean real world how imperfect it is has systems to deal with that.  There's hierarchy of law.  A policeman is allowed to stop your car, right?  So in that instance.  And possibly it's legal upon the country to ask you to pull over to the side of the it's possibly to ask you to pull down your window and look at the occupant, right?  So all of those types of negotiations take place every day.  And they are actually rule‑based in this very Broadway that they use rule‑based.  There's lots of human mockery that goes on.  But there are rule‑based systems.

A subcategory of what you're saying has solutions, right?  Some of the other things need to be negotiated.

>> TRACIE WEISLER:  Who else from our audience here?  Great.  I'm coming over.

>> Thank you very much.  My name is Ismi from Japan.  We are working on same subjects since earlier this year.  And we hold a workshop on the theme in December in January called making the next social ‑‑ through open innovation.  And my question is what the next steps was?  Including is FCC working together with national highway and traffic safety agency?  Because we haven't found that in Japan yet.

While I just read the announcement, Japanese company will open up social field testing of autonomous taxi in three cities in next January.  And so they're trying to really implement this in the hope that it will save our economy more, something like that.

I've been aware of the open source people from Chinese guy, designing the platform whereby the users and the partners who will be able to make cars largely trick, many things are happening.  So the question is:  What's our next steps?  For me, social field testing is inevitable.  Whatever the challenges are.  Just from the table doesn't really make sense to me.  I'd like to hear your views.  Thank you.

>> TRACIE WEISLER:  Before I turn it back to the panel, let me just say I do have a contact at the FCC who is working on a taskforce.  Thanks.  To the panel?  [Inaudible]

>> We have full support.  And we are focusing on public transportation.  The autonomous cars are not focused just on the individual people who buy a car.  It's more for the public transportation.  (alexander speaking).

And in this case, whatever you can do you're going to do.

>> BILL WOODCOCK:  A lot of this will depend on the interaction between regulatory agencies and manufacturers and markets.  So looking at the U.S. which is where most of my experience is, Europe, we have one issue unlike aviation which is federally regulated across the whole country, driving is regulated on a state by state basis.  The laws and regulations are completely different state to state with only a very few not completely uniform harmonization efforts between states.

For instance, up until almost 1930 you could legally drive in one state using a driver's license from another state ‑‑ you couldn't.  You had to be licensed separately to drive a vehicle as well as the vehicle being licensed separately in each state.  We're 80 years on from that now, thankfully.  But you can see how there's a progression towards harmonization over time.

So what we have right now are soft of the early adopter regulators.  In the United States that is Nevada.  That is a big wide open desert state with very little population.  It has a lot of other things that are legal but not legal in other parts of the country.  And so they were the first one to say:  Sure.  Tesla, Google, come on over.  Drive stuff around with no people in it.  We got no problems.  Nothing for you to run into here.

So moving from that to Manhattan is a huge, huge difference in terms of risk and also in terms of regulation.

I think manufacturers are sort of ‑‑ they're hunting.  They're jurisdiction shopping right now.  They're looking for countries, states, whatever that they can work together with where the country and its regulators are eager to do the work in order to be early adopters and are willing to sort of match the manufacturer in terms of work and also accept the risk around an early technology.  And I think that's what you're seeing in Singapore, right?  Singapore has a fairly controlled environment, they figure they can absorb that risk and the advantages that come to them being part of an early adopter system are very high.


>> HELANI GALPAYA:  I think trade is actually going to help in harmonization and Japan really has an opportunity just in terms of ‑‑ because with the exception of India when you think of south Asia, majority is in Japanese cars.  Right now, they do a really poor job of allowing any kind of localisation.  So, for example, you can't select the language to English in your sat nav that's prestored.  Now that's a basic, basic thing that you should be able to do, right?  But think beyond about these systems that allow telemetry and information going back and forth.  If a lot of this stuff is going to be preprogrammed, you're going need very flexible ‑‑ you need to have some flexibility built into the software so at least the basic level of customization can happen when a car lands and goes to the distributor.  I think that will be a starting point.  And then a lot of stuff can be built on top of that, layering of the local customised stuff.

>> TRACIE WEISLER:  Thank you.

>> Yes.  Thank you, Dutch IGF.  I was wondering on the liability part where we see, for example, Google and Volvo and [Inaudible] now say that they take all liability where everything happens with their cars while they're testing.

And then I was wondering how you see that evolving in the future as they say okay it's our software so we are always liable when something happens or if they say no, if you touch the wheel only with your finger in an accident, we will sue you?  So how do you see that evolve in the future?

>> TRACIE WEISLER:  I'm just going to jump in here to say I don't know if we have any lawyers per se on the panel, do we?  No.  But I will say that in the U.S., I'm e not a lawyer, I work with too many of them, actually, the idea of agency, the idea that you have a free will agency decision in your daily life is often a legal framework by which we are judged by our actions.  The basic is almost interesting.  It's almost met an physical.  Maybe Alexander can help with us that.

>> ALEX GAKURUYes.  Car manufacturer side ‑‑

>> ALEXANDER LU:  Car manufacturer side.  What we do, we set up the product.  We make the software in many language.  And we also have a lot of sensors that we collect data and storage, black boxes inside the car.  And these black box can be used as proof for any problems that happens or any accidents.  But we set up for each country we start with.  Regulation, we start out with countries.  And for most complicated, we hold until they get more mature.

>> BILL WOODCOCK:  To number of vehicles are so low and the Google ones are going around 15 miles per hour by and large.  And they weigh 200 pounds or something, right?  They're not going to be serious accidents in any financially interesting quantity in the current iteration.


So credit cards, it's basically universally understood that the holder of the credit card is not liable for fraud that happens on that credit card.  And all of the banks say, "do you know what?  We got this.  Go use your credit card.  Spend a lot of money.  Get a lot of interest.  Everything will be fine.  We self‑insure.  It will not be your problem."

And there is a lot of fraud.  And as people who work on Internet Cybercrime and financial stuff know, a huge amount of that goes unaddressed.  There is data that would allow the apprehension of someone who is doing credit card fraud and it doesn't get used.  The person is not arrested because the bank says "do you know what?  That's under our threshold of caring about it.  We're still within your budget for self‑insurance for this year.  Don't bother."

Now, will automobile accidents be like that?  I kind of doubt it, right?  Because the liability around credit card fraud is typically a few hundreds or a few thousand dollars at most, whereas the damage in an automobile accident is almost always sort of ten times that.  And in the event there's serious loss of life or serious injury, it can be thousands times that.

So the automobile industry globally, I assume to be much smaller than the banking industry globally although they are both juggernauts.  So I suspect that the automobile industry will probably continue to try to go for regulatory regimes and caselaw that will absolve them of a fair amount of the liability or reduce the liability.

>> Alexander:  Just so.  Since we work with public transportation, some of the liability we transfer to the local operator, the taxi companies that manage the fleets, that makes the maintenance, that gets the license from the city, he gets liability because they have operations.  Taxi driver, someone is responsible, anyhow.  We're just hardware manufacturer.

>> Good morning.  My name is mark from State University of São Paulo.  My question relates to the telemetry aspect that was discussed earlier ‑‑ collecting data all the time where they're being operated?  And I have a question, what sort of measures are already being implemented in terms of keeping the safety and the privacy or that's not of concern, what sorts of issues are you seeing right now in terms of that?

>> Alexander:  About the telemetry, the new cars, they collect all the data of the machine when it's working and if it has some bugs, stability, the moment the client charge the cars, they transfer to us.  And basically download the information, we send to collect the car.  This is our plan, especially for resident market.

But for the security, just like when, the specific information, the client must sign an agreement to have this information and we keep as confidential as any company would do.

>> BILL WOODCOCK:  I think you've heard Helani say earlier that this is an area in which consumers are poorly educated until they're not tan they're complacent until they're not, where many countries have very lax regulation and enforcement.  So in general, this is a big problem.  It's an area in which very little will get done until people become outraged and then things will change fast.  And when things change fast, they're often not thought out well.

So education is the way to solve that.  Having people think consciously about these issues and ease into the regulation and so forth.  I was going to point out another example.  Go ahead, I'm sorry.

>> HELANI GALPAYA:  In the absence of governments addressing this and exactly because when the rules come down, they'll come down hard and they will probably be more damaging than good, what we are doing in south Asia is actually working with just the Telcos, not the companies or anything, based on the cell phone privacy issue on how data sharing can be done, under what conditions can and should it be de‑anonymized and shared and when can it be allowed?  And what kind of contracts go along with it?  And we, since about a year ago, we actually got the Telcos to come into a room, put out a draft of core of self‑regulation, essentially, to start with, so we are in version 2 of this.  That's one approach we're trying to take in the absence of the that's not the ideal situation by any means.

>> BILL WOODCOCK:  The thing I was going to say is we've got sort of one collision case coming towards us pretty quickly.  But in vehicles, there's a lot of pressure towards voice control, right?  Because voice is something you can use, control systems, complex systems without taking your hands off the steering wheel, your attention off the road.

We also have consumer devices that implement voice control now.  In order to do this, because the computational power to interpret the voice is not present in the device itself.  And it may not have the expertise to recognize multiple languages, so forth.  That requires realtime streaming of all video environment out where it's centralised computing and handed off to actual people if the compute certain not able to figure it out, right?  That's how you get workable voice control.  If's what the big technology companies have been able to make happen.  And it's likely that that path will continue for quite a while.  There are lots of advantages for dog it that way, technological advantages ‑‑ doing it that way.

Having the entire audio environment of your car being piped out to a computer somewhere completely outside your control so that you can say, oh, I don't like that song, play the next one, it's a big tradeoff.  And it's one people just don't think about right now.

>> TRACIE WEISLER:  Any last questions from our audience today?  Yes.

>> OECD.  I think it is very interesting what Helani raised about the issues of quality of service from wireless networks.  You cannot assure quality of service because it's not the same as in fixed line.  So how reliable is that issue for autonomous vehicles?  What are the challenges there both for operators and for communications authorities?  I mean, should we regulate anything on that?  Is it possible to be done?  Are there any issues to be addressed?  Priority, we would need ‑‑ customization but we may need also regulatory action.


>> BILL WOODCOCK:  That really is the question.  The question is this tradeoff between safety and the vast majority of the SM network.  From an automotive perspective is used for entertainment right now.  People talking to each other, making themselves happy.


Typically required but it needs to happen in very high speeds.  [Inaudible] and the problem is.

[Inaudible] implement in the way delivers on its promise.  If it is possible, it would have worked 20 years ago, right?  People have tried and tried and tried and tried.  The only thing that can work is adding more bandwidth.  The only reasonable thing to do is add more bandwidth.  That is globally standardised that uses different protocols and is not ‑‑ sort of ‑‑ [Inaudible] and different scales, right?  A lot of this is going to be communication with the parking meter that's 2 feet away with the car that's 20 feet in front of you with the car that is swerving into the lane next to you, with the traffic authority that has a broadcast tower 5‑kilometers away.  Not halfway around the world.

So the needs are different.  The solution is probably different if it's done well.  But that requires a huge global regulatory effort.  And we know what that's like.

>> TRACIE WEISLER:  Great.  Well we have about five minutes more.  I wanted to see if the panel had any last words they wanted to share with the group here today?  Any last thoughts?

>> Alexander:  I see here we analyze a lot of possibilities.  For the car manufacturers' side, it is a very tough way because the mind of people is very creative.  We about a lot.  So we need to implement these thoughts we have.  It always takes time.  It is not so easy.  And we need a lot of market support and the government support.

>> BILL WOODCOCK:  We talked a lot about the risks and personal entertainment and whatever.  There are some really fundamental improvements to life that can be achieved through this stuff.  And something Alexander was sharing with me before the meeting is that each of the buses has 300 hours of battery on board.  That's like four teslas worth of battery.  Because this came up on the context of vehicle grid.  It basically is the notion that when the power grid has problems, electric vehicles can supplement ‑‑

(please mute remote participants).


Offduty buses drive themselves to hospital, plug themselves in and power at the hospital.  A megawatt of power, three buses in an hour, right?  So every 20 minutes power hospital.  And then they can drive themselves off somewhere else and recharge.  This kind of thing is going to make disaster recovery response hugely easier, right?  When it can be coordinated and you can have autonomous vehicles transporting people to hospitals that are less crowded further away, you're not going to be limited by ambulances or the ability of power grid not to fall over.  These kinds of things.

As we said earlier, the few your is not distributed evenly.  It doesn't mean that it's not worth doing.

>> TRACIE WEISLER:  We have two more things to do here as moderator.  One is to thank the panelists because I think what Bill did was ended on a high note, the upside.  I think that's important for this panel, as well because we began our hour and a half today with a list of the challenges and the risks potentially and deployment as we move on.  I want to thank Helani because I think that her view from the region was really, really important today.  Her examples were perfect, on spot.  I want to thank Alexander because he gives us a view of how he's dealing with these issues every day in the real world and what it means for the future for his company.  I want to thank S. r Correa for his very brief intervention for Brazil.  I want to have a public service announcement if you will.  The OECD has been instrumental in putting together this panel today.  I'm just a warmup act.  And what I tried to do is bring great people together which clearly we have.  But the real work actually has been ongoing in Paris by the OECD first who I mentioned and then currently Gail Hernandez who is working on a number of interesting IoT aspects in a paper to come out for the OECD ministerial effort which is in 2016 June in Mexico where 34 to 100 Ministers will gather to address these issues.  So we're full of policies.  We've drilled down and tried to narrow them today.  But keep your eye out for OECD work in this area.

And finally I want thank Lorrayne Porciuncula who is from the OECD Secretariat.  She's in the back.  I want to give a round of applause to Lorrayne.  She is actually one of those people in the middle of the night connected, not by vehicle but by her Android to the speakers tonight from last night to today and trying to get all of the remote possibilities handled.  So thank you so much all of you.  Appreciate your questions.  And think about your next car.  Bye‑bye.


[End of session.]