Christopher Hart: Opportunities and challenges of autonomous vehicles

February 10, 2020 1:15:12
Kaltura Video

Christopher Hart describes the policy opportunities and challenges that autonomous vehicles present policymakers in 2020. February, 2020.

Transcript:

Welcome.

My my name is robert.

I am an ashirr yacht professor 

At ford school of policy.

I have affiliations with the 

University of michigan 

Transportation research 

Institute.

So, welcome today.

It's may pleasure to introduce 

Tower speaker and our lecturer 

Today on the opportunities and 

Challenges autonomous vehicles: 

The roles for government f. This

Is sponsored by the center for 

Local, state policy and the ford

School's public policy talk 

Series.

Also, I'd like to thank our 

Co-sponsors today, which are the

Science and technology and 

Public policy program and the 

University of michigan 

Transportation research 

Institute.

So, thank you for all the 

Sponsors and co-sponsors.

Bench ever been are -- we'll 

Take questions at the end from 

The audience.

So, you will see people who have

Index cards.

Please write your questions on 

Those index cards and we'll 

Start to collect those around 


Then ask those questions.

So, today we have some ford 

School students who are going to

Help ask the questions.

So they will help ask the 

Questions also with support from

Christy richardson.  So thank 

You. So let me introduce our 

Speaker today, Christopher Hart,

Who is the former chairman of 

The national transportation 

Safety board.

He has the honor of being 

Appointed by two u.S. Presidents

By different parties.

George bush senior and also 

President obama -- to serve on 

The board.

Chris is an engineer but also a 

Lawyer who worked at the federal

Aviation administration for 

Years, at nitsa for years and in

Private practice.

I had the pleasure of serving --

Being on a panel with him last 

Year at princeton reunions.

We have both princeton 

Backgrounds.

That was a great time.

So I figured the ford school 

Community could benefit from 

Hearing his sage advice based on

His years of experience in his 

Career.

At this point I'd like to 

Introduce Christopher Hart and 

Take it away.

.

Christopher Hart.

Good afternoon everyone and 

Thank you very much robert for 

That kind introduction.

This is an honor to be here and 

Have a chance to talk to you 

Guys about something that is 

Going to be transformative to 

Everything we do and that is 

Autonomous vehicles.

The specific question is what 

The role of the governments in 

This equation.

In order to do that I have to 

Lay the ground work to some of 

The benefits and challenges of 

Autonomous vehicles which I will

Do.

So, thank you for the 

Opportunity to be here.

It is a privilege and honor and 

I'm going to -- if robert has 

Anything to say about it, 

Hopefully I'm going to say 

Something that you guys enjoy 

Hearing that will be productive 

To you.

I'm going to talk about the 

Opportunities, talk about the 

Aviation mode that's most 

Advanced in automation.

And that's aviation and lessons 

Learned from them.

There are quite a few other 

Lessons about automation that 

The carmakers are going to have 

To address because automation on

The ground is much more 

Challenging and complex than 

Automation in the air.

There's going to be a lot of 

Other challenges for the car 

Community that weren't faced and

That won't be faced in aviation.

Then I will end it up with the 

Roles of the government.

So the opportunities are that 

Automation has been a huge 

Success story everywhere that 

We've seen it at the national 

Transportation safety board.

By the way, let me give a little

Background.

The ntsb is the investigative 

Agency for the federal 

Government.

They investigate accidents in 

All modes of transportation and 

Come up with findings and 

Recommendations and the 

Recommendations are intended to 

Help prevent recurrences of 

Those accidents.

Let me just emphasize, they are 

Just recommendations.

The ntsb can't require anybody 

To do anything.

Contrary to what the media would

Have you think, well when I 

Receive this letter from the 

Ntsb that says we recommend such

And such -- I emphasize the word

"Recommend" -- so I don't have 

To did it.

More than 80 percent of the time

The recommendations of the ntsb 

Are followed because they're 

Good ideas and if you don't 

Follow that recommendation then 

Have an accident because you 

Didn't then you really have some

Explaining to do to the jury 

After that.

So that's really the main 

Leverage for the ntsb and why 

More than 80% of the time even 

Though people don't have to they

Still do the recommendation or 

Something like it to help make 

Transportation safer.

That's what the ntsb does.

And they do it in all modes.

What really got me interested in

The highway mode, because the 

Ntsb didn't do historically that

Many highway accidents, and most

Of them that we did were big 

Vehicles, buses and trucks with 

Commercial/professional drivers 

As opposed to joe public with 

His car.

And this one, the one I am going

To talk about, which is the 

Tesla accident in florida in 


Had seen in my ntsb career that 

Involved joe public and his car.

And we looked at that because of

The automation impacts because 

We had been dealing with 

Automation for years and years 

Especially in aviation but also 

In other modes of 

Transportation.

In aviation automation has shown

Amazing safety benefits, 

Productivity and operating 

Efficiency, getting more 

Airplanes through the airspace 

And reducing pollution at the 

Same time.

I think the car automation 

Scenario has even more 

Opportunities for improvement, 

Especially in safety and many 

Others.

So the question is how do we get

From here to there.

Because I am a gadget geek.

What I say you may think I'm 

Anti-automation but I'm pro 

Automation.

I am a gadget geek.

The question is, how do we get 

There without hurting too many 

People in the process.

Because the public is already 

Skeptical of the whole notion of

Ought mace, including me.

Over time they see an accident, 

I have to wonder, yet again do 

We see the car people not paying

Attention to the aviation 

Mistations that have occurred 

And making the same mistakes 

Over again which is a shame 

Because it makes the public even

More skeptical.

How do we get from here to 

There?

I am convinced we will be there.

The estimations of it were 

Wildly over optimistic.

Because it wasn't that long ago 

That estimates were by 2020 the 

Streets will be full of 

Driverless cars.

It's 2020 and guess what.

That's what I am going to talk 

About is what the role of the 

Favorite governments -- federal,

State and local -- in helping to

Make this happen.

Let's look at the lessons 

Learned in aviation.

When they first started 

Automating in aviation they said

Here we have the technology to 

Do it, let's do it.

That led to bad results.

So that's when they realized 

Maybe we ought to focus on human

Sent rick automation.

Which means developing the 

Automation around the human.

When automation isn't perfect in

Cars you need to have graceful 

Exits.

You need to have a graceful exit

If the driver is not paying 

Enough attention.

The florida accident is one 

Where the driver wasn't paying 

Attention.

In addition, you got two big 

What-ifs.

What if it fails and it's not 

Doing that very much anymore.

The automation is very reliable.

Number 2, what if it encounters 

Unanticipated circumstances.

Circumstances that weren't 

Anticipated even by the 

Designers of the automation.

That's when you need graceful 

Exits.

That's when people ask me:  how 

Long before I get on an airliner

With no pilot?

No time soon.

Because until they figure out 

The what-ifs, what if the 

Automation fails or encounters 

An anticipated circumstance, 

I.E. Sullyly, until that happens

We're not going on see airplanes

Without pilots.

Last but not least another 

Lesson is when you have a system

That's very reliable and you ask

A human to be a monitor for that

System, that doesn't work very 

Well.

Human-centric, as I said 

Automation became with because 

We have the technology do it.

Now they're human centric and 

The carmakers are trying to but 

The ongoing crashes reveal they 

Need to do better human-centrif.

To me the word auto pilot is a 

Huge mistake.

I am a pilot and I know when the

Auto pilot is on I'm still 

Important.

I am not going to go to sleep.

What joe and susie public hear 

Auto pilot they don't know that.

Auto pilot gives the impression 

I will do this and I am not 

Necessarily here.

When the systems are very good 

Then if your system is really 

Good you are going to encourage 

The driver to disengage.

Example is I recently got a 

Volvo and my volvo has lane 

Keeping assist.

So the assist is kind of a 

Little sloppy so it will drift 

To one side and hit that line.

By the time you do that a few 

Times that is annoying so you 

Hold the steering wheel so you 

Don't do the back and forth.

Tesla has tight gains so you 

Don't do the back and forth.

And that engages the driver to 

Disengage.

To me that's another failure to 

Learn from aviation automation.

Another one is what's the 

Graceful exit if the driver is 

Not attentive.

The two I've seen is, turn the 

Automation off and stop the car 

In the road.

I am not satisfied with that 

Because what if the person has a

Medical event and you turn it 

Off now you are asking for a 

Crash.

And stopping the car on the 

Road, you can't tell you how 

Were of many crashes we've seen 

Because of stopping on the 

Interstate.

The one you might remember was 

The one on the jersey turnpike 

Where there was a construction 

Cue stop and the walmart truck 

Ranked into the construction cue

And killed tracy morgan's buddy 

And seriously injured tracy 

Morgan.

Usually you have two miles of 

Signs, warning, construction 

Ahead, et cetera.

If a car just stops -- and still

You had the truck coming in at 

Highway speed.

If you don't have those two 

Signs, what are we going to see 

For people stopping on the road 

Because most of the systems 

Today say don't use them except 

On interstate highways.

Well if you use it on interstate

Highway and the car stops in the

Highway that's pretty much 

Asking for an accident.

So, to me we're not ready 

Because we don't have a graceful

Exit for drivers inattention.

If the automation is not 

User-friendly, and much isn't, 

Then it can be used improperly 

Or distract the driver.

What is that chime I am hearing?

You are looking for the crash 

Then you run into something 

Because you weren't looking at 

What's in front of you.

None of those are desirable 

Outcomes.

What about automation being 

Uncertain because the lane 

Markings after the sand and 

Salt, that's going to wear them 

Down.

What happens when the lane 

Markings become not so good or 

In my car, again, with the lane 

Keeping assist, when you come to

The exit, then the right lane 

Now die verges from the traffic 

And the car starts to follow 

That until it rights I don't 

Want to do that so that's just 

Another example of what about 

Uncertainty because of the 

Markings or because of the 

Streets are slippery.

What's the story on will the car

Affect its stopping distance 

Algorithm.

What about unanticipated 

Circumstances like an object in 

The road.

I'll be talking about tempe, 

Arizona where that happened.

What if the system is uncertain.

Does it warn the driver in 

Enough time for the driver to 

Take over and handle the 

Situation?

So here's williston, florida, 

That I talked about.

This is where a tractor-trailer 

Opposite direction was turning 

Left at an intersection on a 

High-speed road but it was not 

An interstate it had occasional 

Intersections.

The tesla submarined under the 

Trailer and, of course, it 

Decapitated the driver by 

Sheering off the top of the car.


Marks.

This is one where the driver had

Been bragging about I got my 

Auto pilot so while it's driving

I'm playing sudoku.

Well the owner's manual, tesla's

Said, that shouldn't have 

Happened because the owner's 

Manual says only use it on 

Highways and limited access 

Roads.

Well two problems with that.

One is, I think this was a 

Highway.

To me the person was following 

The manual.

Number 2, who reads the owner's 

Manual?

I read it twice a year.

Daylight savings time starts and

Daylight savings time stops, how

Do I change the clock.

These days you don't even have 

To do that because the clock 

Changes itself. So much for the 

Two times a year when I look at 

The owner's manual.

So the point is tesla was over 

Relying on the driver looking at

The owner's manual and the 

Driver was over reliant on 

Tesla's ability to develop 

Automation that didn't need him.

As you can see it barely touched

The truck but the driver was 

Killed by sheering off the top 

Of the car.

Here is another one, and this 

Happened to me in a regular car 

Where I'm behind a car and the 

Car suddenly moves out of the 

Lane and I realize they moved 

Out of the lane because there 

Was an obstacle in front of that

Car.

I am thinking I was lucky I 

Could move out.

If someone had been besides me I

Would have run into the object 

In front of us.

The tesla was behind a car in 

The left lane, there was a fire 

Truck in the left lane dealing 

With an accident.

The car moves to get out of the 

Fire truck and the tesla 

Continues on and crashes into 

The fire truck.

This is one I think I may have 

To update this slide.

I think the ntsb may have 

Completed the investigation.

I'll have to see.

But the point is this was one 

Where once again the auto pilot 

Was not up to the task and you 

Need the driver to do the job.

Here is where I was looking for 

A situation where the automation

Failed -- and that is very rare 

Because the systems are reliable

These days -- I couldn't find 

Any in aviation.

I had to look to the subway 

System in d.C. To find a 

Situation with automation actual

Failure.

This is one where the trains are

Registered on the system 

Electronically and occasionally 

For reasons we didn't know 

Before the accident the 

Electronic signature disappears 

And when the electronic 

Signature disappears we know in 

The main dispatch board that 

Shows where all the trains are.

When it disappears they get a 

Warning at the dispatch board.

We asked them what do you do 

With that?

They said nothing because we get

That warning 500 times a day.

So that's not very useful.

Guess who didn't get the 

Warning?

The train behind the train that 

Stopped because this train that 

Disappeared had stopped and the 

Train behind it didn't get the 

Warning so the train is saying 

The track is empty we'll 

Accelerate to full speed.

So that's what the train was 

Doing when it rounded the curve.

That's what limited the sight 

Distance of the operator.

As soon as he saw the train 

Stopped on the curb, hit the 

Emergency brake but it was too 

Late and it killed her and eight

Passengers.

So if the automation fails and 

It doesn't warn the operator in 

Time, you got a huge problem.

And that's one of the issues for

Car automation.

When it's uncertain or fails, 

Does it warn the driver in time 

To let the driver be able to 

Salvage the situation?

Then there's what I say happens 

Much, much more often.

I could give you 15 to 20 

Aviation accidents involving 

Unanticipated circumstances.

But this is sort of the really 

Example of how even aviation 

After automating for decades 

Doesn't have it totally figured 

Out.

So this was air france from rio 

To paris.

As most flights are from north 

And south america to europe it 

Was a red eye at night.

They're taking off and go to 


And they're near thunderstorms.

As soon as they reach cruise 

Altitude the chief pilot who was

The most experienced on board 

Goes back to take a scheduled 

Rest break.

So, question number one is:  was

That a good time to take a 

Scheduled rest break when you 

Saw on the radar there were 

Thunderstorms 80 to 100 miles 

Ahead of you.

Here's all this super cooled 

Water because of all these 

Thunderstorms out there and the 

Airplane has what's called a 

Peto tube which sticks forward 

And has a hole in the front.

The purpose of the hole is the 

Air goes in and the airplane 

Knows how fast it's going by how

Hard air is coming in.

You want to make sure the hole 

Doesn't freeze over because then

No air comes in and the plane 

Has no clue how fast it's going.

To prevent that from happens you

Have heaters on the peto tubes.

Well a couple of times this 

Particular type of airplane had 

The heaters overwhelmed by the 

Super cooled water and it froze 

Over the tubes and the pilots 

Pulled out of the situation 

Successfully.

They figured we probably need 

Better heaters on the peto tube.

Number 2, they pulled out okay 

So it's probably not an 

Emergency.

We'll just do it the next time 

This airplane has regularly 

Scheduled maintenance which this

Plane was scheduled to do in the

Next two or three weeks.

So, now when the ice blocks the 

Peto tubes and the airplane 

Doesn't know how fast it's going

There are lots of systems that 

Don't operate if they don't know

How fast it is going.

For example, the automatic 

Pilot, throttle, it won't 

Operate.

The protection against stall 

Won't operate.

So, there's lots of systems.

What the pilots would have 

Gotten are a bunch of air 

Messages that you have lost your

Airspeed information but they 

Would also have eight or six 

Messages of the systems that 

Quit because they don't have the

Airspeed information.

So the pilots responded 

Inappropriately and ended up 

Crashing.

In that one, we looked at the 

Totality of the circumstances as

We always do.

When I say "We" not the ntsb 

Because we weren't in charge the

French aeration of the ntsb was 

In charge.

But did the pilots ever have 

Training about what happens when

You lose air information in 

Cruise.

No.

They never had training about 

That.

Did they have training about 

Flying manually at cruise 

Altitude?

Most of the time they don't.

Most of the time it's illegal to

Fly manually at cruise altitude 

Because back in the old days 

They use to have 2,000 feet 

Between option direction 

Traffic.

When they needed more highways 

In the sky they reduced to that 


The designer of that system said


Away.

There's too much of an 

Opportunity of a mid-air 

Collision because 1,000 feet, I 

Don't trust humans to to do that

Successfully so I require you be

On automation.

Anything above 29,000 feet you 

Have to be on automation.

That's why they never had 

Training about how to fly at 

Cruise altitude and it is a very

Different airplane.

Have they recovered from a stall

At cruise altaltitude?

Definitely not.

So they hadn't had that.

Plus they weren't talking to 

Each other.

So, as soon as the airplane lost

The airspeed information the 

Right side pilot yanks back his 

Side stick which causes the 

Airplane to start to climb but 

He didn't talk to the left side 

Pilot to tell the pilot what he 

Was doing and another link in 

The chain is if this had been a 

Boeing airplane, and I am not 

Slamming airbus, but this is the

Difference.

If it was a boeing airplane 

Where you have a control yolk 

Between your legs if the 

Right-side pilot pulled it back 

To his belly it would be in my 

Belly too because they are 

Connected.

But when you move the slide 

Stick the left side slide stick 

Didn't move.

So, he had no clue.

They weren't talking to each 

Other to make sure they're on 

The same page.

Bottom line, they responded 

Inappropriately and crashed.

So this was just a poster child 

Example of how whole system not 

Having their hands fully around 

Automation.

So, those are some of the 

Challenges that the car industry

Is going to have to pay 

Attention to that aviation has 

Faced.

There's a whole bunch of others 

And to save time for answering 

Questions I won't governor these

In detail but I am going to over

Them a few words.

Artificial intelligence, learns 

With experience, drivers not 

Trains, street testing 

Essential.

I'm going to talk about how 

These are additional problems.

Automation on the ground is much

More challenging than automation

In the air.

Then I'll talk about what roles 

The government is going to play 

In all these situations.

Automation that learns.

Doesn't happen in aviation 

Because whenever the aviation 

Changes in aviation you have to 

Retrain the pilots to the new 

Automation.

That's one of the questions 

That's come up in this 737 

Situation which I am also 

Working on.

So, in aviation, if the 

Automation changes enough, they 

Have to retrain the pilots.

Once the automation is in place 

It's locked down and that's what

They train the pilots to do.

And the pilot do's that until 

The automation is changed again 

And then it's locked down.

There's no artificial 

Intelligence or learning in 

Process while the automation is 

Working.

So that's going to -- to have 

That automation that changes so 

When you come out to your car 

The next morning and then you 

Get the word there was a 

Software upgrade you wonder if I

Am going to be able to handle 

This car the way it behaves now 

Versus yesterday when I left it 

And how much training is there 

Of drivers?

Zero.

So, there's not going to be any 

Training of drivers.

That's going to be a huge 

Challenge.

And that means when the 

Automakers change the automation

They have to assume two things.

One is there won't be any 

Training because that is never 

Going to happen.

Number two is that the owner 

Won't read the owner's manual 

And, of course, the change from 

Last night wouldn't be in your 

Manual anyway.

You have to assume the worst 

Case that how driver friendly is

This for the driver to respond 

To this new change in 

Automation.

So that's an example of an issue

They're going to face on the 

Ground they don't face in the 

Air.

Airline pilots train and retrain

Regularly.

It doesn't happen in cars.  So 

That means designers of 

Automation have to assume worst 

Case.

Drivers don't have training and 

Won't look at the owner's 

Manual.

So the possible outcome.

You turn the automation off and 

Lose the protection.

And, b, you become distractioned

Then you have a crash because 

You are distracted.

The question for the auto 

Dealers is -- and I gave this 

Presentation to the dealers 

Once -- I said today the dealer 

Will just hand you the key.

Is this going to change that 

Liability scenario so they have 

To not only give you the key but

Train to you some extent about 

How to use the car.

I am not just talking about 

Pointing you to a website but 

Training you about how the car 

Works.

So, stay tuned on whether that's

Going to change.

How many people in the auto 

Showroom do you know of that 

Substantively familiar enough 

With the car to train you?

So that would be a whole 

Different experience going to an

Auto dealer and having someone 

Trained enough to tell you how 

It works.

This is one you will be 

Interested in because you got 

The m city, which is your test 

Track, but what we're finding 

Big time is actual street 

Testing is essential.

No matter how much test track or

Lab testing you do, actual 

Driving on the streets is 

Essential to make a car street 

Ready.

So that's going to be a big 

Challenge.

The reason that is a big 

Challenge is because it raises 

The essential conundrum assuming

You are a responsible enough 

Manufacture that you do test 

Track training and lab training 

Before you put your car on the 

Streets that means you have a 

Reliable car.

Now the conundrum is what I said

About aviation.

That humans are not good 

Monitors of reliable systems and

That's what happened in tifty.

If it's not sure what's going 

On, it needs to warn you in a 

Timely manner.

Also it needs better monitor 

Training.

To tell the monitors this is a 

Scenario that humans don't do 

Very well.

Namely monitoring reliable 

Systems.

And here is how we're going to 

Train you.

So, tempe, arizona was a 

Driverless street test with the 

Monitor.

You have seen the youtube where 

The monitor was looking down at 

The moment.

To the woman was walking across 

The street at night not in a 

Crosswalk.

I think the ntsb investigation 

Of this one might be completed 

As well.

But this was the first fatality 

Of a pedestrian from a, quote, 

Driverless vehicle.

This is one again where the 

Driver wasn't paying adequate 

Attention.

But look at the way this works.

It doesn't matter what the 

Driver was looking at.

The driver could be distracted 

By a whole bunch of things many 

Of which are legitimate 

Distractions.  I could be 

Looking at the street sign or 

The rearview mirror.

I did have an accident where I 

Was looking in the mirror and 

The guy in front of me stopped 

At that moment and I crashed.

So, there's no way that even 

With a driver that's not 

Distracted by things they 

Shouldn't be distracted by like 

Their ipad or iphone, there's 

Still no guarantee the driver is

Going to be 100 percent percent 

Of the time looking out.

That's why this accident was 

Pretty much inevitable.

With all the street testing 

Going on that accident was 

Pretty much inevitable.

I was engaged by the law firm 

That was asked after this group 

Uber stopped testing and they 

Engaged a law firm to get them 

Back on the streets and the law 

Firm asked me to help them.

I that have started on the 

Streets in pittsburgh.

They did start back on the 

Streets.

But the point was that this is 

An example where it was 

Inevitable with all the testing 

Going on there way is going to 

Be a moment in time when the 

Person was not looking out at 

The moment they needed to be.

There was an structure issue 

Because the sidewalk didn't 

Have -- it stopped at the street

And didn't have a crosswalk.

So this woman was walking on the

Sidewalk then crossed the 

Street.

By the way she was high but 

That's something neither here or

There.

So, there was a lot of criticism

About how come uber turned off 

The volvo driver alert system to

Keep collisions from happening.

The reason they did is because 

They wanted to invent a system 

That would work on any car not 

Just top of the volvo system.

So they disabled the volvo 

System to detect objects and 

Stop in time.

This is another example of 

Whereas we try to figure out how

To do this better, there's going

To be chanllenges.

This is people trying to do what

The think is right in that 

Moment but this is going to 

Maker take, again, better 

Warning of uncertainty and 

Training of the monitors to say 

We're asking to you do what we 

Know humans don't do very well 

Which is monitor reliable 

Systems.

Graceful exits, as I mentionth 

You got to have graceful exits 

If the automation fails or 

Encounters unanticipated 

Circumstances.

That's important when you have 

Drivers in the car.

If you take the driver out of 

The car it's crucial.

It's not just important it's 

Essential.

Graceful exit.

This is like I say when people 

Ask how long before I get on an 

Airliner with no pilot, the joke

About the airliner in the future

Is you look up front and there 

Will be a pilot and a dog and 

The pilot's job is to feed the 

Dog.

The dog's job is to bite the 

Pilot if he touches anything.

Here's one where both engines 

Ingested birds.

Pilots were unable to reach the 

Airport so they landed in the 

River.

What would automation have done 

In that situation?

When I talk to aeronautical 

Crowds they tell them there was 

Actually automation undermined 

This situation.

Because this was one where sully

Tried to do asr as soft and 

La ing as possible by flaring 

Just before he hit the water to 

Minimize the vertical impact 

Speed of the airplane on the 

Water.

The automation that was on at 

The time which is called a fugue

Odd damper, kept him from 

Pulling the nose up as much as 

He wanted to, stopped him from 

Pulling the last 

Three-and-a-half degrees of nose

Up which caused him to hit the 

Water harder than he anticipated

Which breaches the bottom of the

Fus

Fuselage.

So, question was, if that 

Automation had not stopped him 

From doing that -- by the way 

This is automation he had no 

Clue of and he was very upset 

When he found out there was 

Automation that kept you from 

Flaring as softly as you wanted 

To, he was very upset as any 

Professional would be.

Because any professional wants 

To know what does the airplane 

Do, what are the capabilities of

This automation.

What does it do and not do.

So, he hit the water much harder

Than he intended to.

That was one where automation 

Undermined the outcome that he 

Wanted.

Mixing driverless with humans.

Again, very challenging because 

Of the lack of variability of 

The humans out there.

So, even if you design a 

Driverless car, you have to 

Consider the other humans in the

System like the pedestrians, 

Bicyclist and the motorcyclists 

And people who you know who you 

Will take this steering wheel 

Out of my cold dead hand, it's 

Much more challenging to mix 

Automation with humans than just

To have straight automation.

Software updates.

That's what I said about are you

Going to be able to handle the 

Difference of behavior in core 

Car the next morning.

Look at how thoroughly complex 

Software is tested for 

Unintended circumstances.

Sometimes I wonder with the 

Frequency of these updates how 

Thorough is that testing to see 

If there are any consequences 

From adding the new software to 

The existing system.

I have a feelings there not 

Adequate testing of that before 

They bring the new systems in to

Place.

Aviation system designers figure

Out when they change the 

Software they've figured out by 

Bringing pilots in to fly they 

Just don't rely on human factors

They bring the pilots in to fly 

In the simulator before it goes 

Out to service.

That's not so easy in cars ba us

The simulators aren't as good 

And there's a much greater 

Variability of drivers.

So that's going to be much more 

Challenging in cars than it is 

With airplanes.

Cyber protection.

That one really worries me 

Because I don't see much about 

It with manufactures.

I see a lot of academia interest

In it like here, but that's kind

Of scary because not only do you

Have to be protected against 

Today's cyber invasion protocols

But today's cyber invasion 

Protocols are ever advancing so 

It's got to be a continuous 

Thing am you have to get better 

Against the advancing protocols.

And I'm concerned I don't see 

How well that is happening.

Not only that, but every time 

You add additional protections 

To your software to protect it 

From cyber attacks, then you got

To, again, just as with any 

Other software change you got to

Look at the unintended 

Consequences of the change.

So, I'm very concerned that 

Inadequate attention is being 

Paid to the whole area of cyber 

Protection.

This is an interesting one.

Competition regarding safety.

Airlines don't complete on 

Safety.

They have a cultural aspect that

Not very many industries have.

That is, anybody's crash is 

Everybody's crash.

So, when public sees an air 

Plane crash, they don't say that

Airplane was airline x and I'm 

Not worried because I am going 

On airline y.

They say wait a minute that was 

A german airplane in that 

Crashed in the french alps -- 

That was a suicide by the way --

But they go I'm worried because 

I'm flying to pittsburgh.

The only other industry that is 

Like that big time is the 

Nuclear power energy.

They are very much that way.

Anybody's accident is 

Everybody's accident.

So the air lines don't compete 

On safety.

That's why you never see an ad 

That says we are the safest 

Airline out there.

They scratch each other's back.

Not only do I not want to crash 

Myself I don't want roberts 

Airplanes to crash either.

Anybody's crash is everybody's 

Crash.

On the other hand, the 

Automakers compete on safety.

So, me that's good.

To the extent you buy the car 

With the five instead of four 

Stars, that helps safety 

Innovations penetrate the fleet 

Faster.

I'm not sure it would a good 

Idea to transfer that model from

The aviation system to the car 

System and say you shouldn't 

Complete on safety.

The trick is how do you take 

Advantage of the fact that it 

Penetrates the fleet faster when

It's got five stars and still 

Have opportunities for the 

Safety messages to be 

Transmitted to everybody so 

Everybody knows how that works.

So that's a challenge.

I think there's a way to take 

Advantage of that safety benefit

Without reducing that system 

That now helps safety 

Implementations penetrate the 

Fleet faster.

We don't see any ethical issues 

In any of the automation in any 

Other mode before today.

The example I've often used is 

You are going along in your lane

Fat, dumb and happy and for 

Reasons that don't matter 

There's an 80,000 truck coming 

At you.


So what is your automation going

To do?

Take you and into the truck to 

Protect the pedestrians or take 

Out the pedestrians to protect 

You?

That's the kind of issue that I 

Don't think should be decided 

After the fact.

I think we need to have serious 

Discussion.

But I don't see any talk about 

Ethics except possibly in 

Academia.

The sooner you put ethic issues 

Into the system the better.

It's just like safety.

If you put it on after the fact 

Its ace not as efficient and 

Effective.

I think the same is true with 

Ethical protections. How are we 

Going to address the ethical 

Issues faced by this automation 

Mode?

I don't have a good answer 

Except I think the feds need to 

Get seriously involved in this 

One.

So, let's look at some of the 

Roles of the government.

I am going to start with the 

Federal government because I 

Think it's very important for 

Nitsa to stand up to this 

Challenge which they're not 

Doing now because they're in a 

Regulatory regime that says if 

You want one new regulation you 

Got to get rid of two old 

Regulations.

That is a pretty arbitrary 

Guideline and puts a constraint 

On what they can do.

I think it's important for two 

Reasons to have the feds get in 

Charge.

One because you don't want 

Different requirements in each 

State with a patchwork quilt of 

Requirements from state to 

State.

The other is the fact there are 

Lots of countries involved.

This is a worldwide issue.

Japan, germany, sweden and korea

And all the countries that make 

Cars need to be engaged in this 

As well.

Who is that going to be with?

Not with the state of michigan 

But with the federal government.

They need to be engaged to 

Effectively do that.

So I think there's a huge 

Opportunity, a huge requirement 

Need for the federal government 

To become more engaged in this 

Endeavor and I don't see it with

Today's federal government.

Let's move to the state and 

Local governmentish yous.

If the feds -- to the extent the

Feds don't establish the 

Requirements, the states will 

Have to do it.

What about licensing 

Requirements?

Which the states have done.

What kind of license are you 

Going to need for a driverless 

Car?

What about hands the wheel?

Do you have to have hands on the

Wheel.

Some states say yes and some 

Don't.

About the requirements having a 

Steering wheel or brakes?

What about the infrastructure?

How much is that going to be?

What do the states and cities do

Regarding the infrastructure?

Are we still going to have 

Traffic signals, lines on the 

Streets?

What about street signage in 

General?

What about street parking?

Predictions are there's going to

Be fewer cars because we're 

Going to move to a system where 

People won't own cars but 

They'll beckon cars.

That means instead of being used

One or two hours a day like your

Car is they are being used much 

More which means you are not 

Going to have -- you are going 

To have a lot fewer cars.

And you are probably not going 

To own your own car because it's

Not worth it for you to put all 

The fixed costs in to buying an 

Asset that's going to be parked 


So, there's a whole lot of 

Changes.

But garages what about the 

Revenue the cities and state get

From the parking garages?

Should there be dedicated lanes 

For av's.

How about bicyclists and 

Pedestrians, should they be 

Segregated?

There's a lot of issues they 

Have to address.

What about the revenue from auto

Registration with fewer autos?

What about the fuel tax because 

If cars are going electric 

What's that going to do to fuel 

Taxes?

To all the revenues the cities 

Get from parking and moving 

Violations?

What about parking lot taxes?

I mean, there are so many issues

That nobody really has the 

Answer to because this is going 

To be so transformative until we

Get some idea what direction 

This is taking it's going to be 

Very difficult for the states 

And cities to respond in a way 

That helps to compliment these 

Changes because we don't know 

Yet what direction the changes 

Are going to take.

What about resources?

Will the need as many traffic 

Police?

Will they need as many ambulance

Services?

What about infrastructure 

Maintenance?

All these issues -- what about 

Slippery streets?

Again the automation today is 

Not able to handle slippery 

Streets.

Here's one that is not a state 

Or local government issue but 

What about organ donors?

If we are not killing 100 people

A day like we are today, what's 

That going to do to the organ 

Donor situation?

Stay tuned.

So, automation offers lots and 

Lots of potential benefits and I

Am a big fan of automation and 

It's going -- not only the lives

Saved but so many other benefits

But there are lots of 

Challenges.

First the auto world needs to 

Pay attention to the aviation 

World which has been automated 

For decade.

It is a shame every time we see 

A lesson learned that's not 

Followed by the auto industry 

And they make the skeptical 

Public even more skeptical.

Not only do they have those 

Issues but a whole bunch of 

Issues because automation on the

Ground is much more challenging 

And complex than in the air.

There are a bunch of issues the 

Auto world will have to deal 

With.

Anticipate big changes for 

Everybody.

This will be transformative.

If you tried to predict ten 

Years ago you'll the things your

Iphone will do now you would 

Probably be way off.

Nobody -- I've seen so many 

Predictions on how this is going

To look that nobody is really 

Ready, none of the states and 

Local governments are ready to 

Put veer use infrastructure 

Changes in place because nobody 

Knows which way this is going to

Go.

Thanks again for the opportunity

To come and talk.

This has been fascinating and I 

Look forward to the questions 

That you might have.

Thank you very much.

[applause]

moderator: thanks, everyone.

Now we have questions.

So, our illustrious ford school 

Panelists will ask those 

Questions.

Also, we can take questions via 

Twitter, the hashtag #policy 

Talks.

thank you so much for your 

Time.

I am a second year mba focused 

On emerging tech regulation.

Interested in the disruption 

Happening in the auto industry.

So, really excited to hear you 

Speak today.

The first question here is:  are

The auto companies aware of the 

Lessons you have described?

Christopher Hart:  that's a 

Real good question.

I don't see them following it 

But there's no secret.

These are in the public domain.

It's not like they have to find 

Somebody who is going to shoot 

Them if they tell them these 

Secrets.

This is out there in the public.

If they are aware of it, I'm not

Seeing it because to the kent 

They are making the same 

Mistakes over again.

Either they're not aware or 

They're not paying attention 

Because they think that's 

Aviation and it's not 

Applicable.

are they invested in learning

From these lessons?

What do you expect from the 

Future as you talk about the 

Changes we should expect.

Christopher Hart:  I'm hoping

If I give this presentation in 

Enough places I'll capture 

Attention to make them realize 

There's lots of history for 

Automation history development 

In aviation and they need to pay

More attention to the history.

thank you.

I am a second year at the ford 

School and the co-chair of 

Mobility policy lab, a student 

Organization focused on growing 

Discussion of mobility and 

Transportation policy here at 

Ford and across campus.

We have several questions about 

States.

I'm going to try to condense 

Them into one.

What challenges do you see in 

Lieu of unified federal 

Regulation about competition 

Between states in developing 

Smart policies?

Where are the potential fault 

Lines, where are the incentives,

The drivers that could produce 

Error and other challenges?

Christopher Hart:  the best I

Can say on that since I am not 

Invested in this industry 

Because most of my work was 

Aviation based, so I'm just 

Hoping that the states are 

Talking to each other.

I don't know if there's a trade 

Association they can work 

Through.

But they're talking to each 

Other so that will create some 

Level of harmony between the 

States.

So, I don't know how much that's

Happening.

I just hope it is.

Because to me, 50 different 

Answers not going to be a good 

Solution.

do you think that the states 

Currently have the appropriate 

Resources to be able to handle 

Such complex systems in a way 

That both ensures safety in 

Consistency but also meets the 

Policy outcomes that states are 

Interested?

without knowing what kind of 

Changes they need to make and, 

Therefore, we don't know whether

Some of the changes are going to

Reduce expenses here and just 

What the nature of those changes

Are going to be, that's very 

Difficult to know at this early 

Stage.

It's going to be -- an 

Interactive situation where the 

States and local governments do 

Is going to depend on the 

Carmakers do and vice versa.

So it's very difficult to 

Predict where that's going to 

Go.

I'm sorry not to have the answer

But this is so transformative 

That I think it's going to be 

Much bigger than not just having

Parking on the street but a much

Bigger change than that.

so switching gears a little.

Ha ha.

Okay.

No?

[laughter]

Sorry.

Christopher Hart:  electronic

Cars won't have gears 

that's probably why nobody 

Laughed.

The chris ler's 3,002,020 model 

Has no printed owner's manual 

And it's on the app console.

How can a car built that way 

Adapt to being an autonomous 

Vehicle.

Christopher Hart:  well the 

Cars that do have manuals nobody

Reads them anyway.

But I wonder -- ask me again the

Question.

I'm not sure if I understood.

it's basically the crux is 

How does a car built that way, 

That is basic with no printed 

Owner's manual, so 

Technologically advanced how 

Does it adapt to being an 

Autonomous vehicle.

I think the crux is whether it's

An iter active process within 

The software or how that might 

Change in the future.

the only thing I can say is 

The way the owner's manual is is

Probably ill rear vanity because

Nobody looks at it anyway.

If it's electronic owner's 

Manual they can adapt as the car

Is adaptsing and that's a good 

Sign.

But to the extent nobody reads 

It.

we have a slough of questions

About insurance.

Could you discuss briefly what 

Role insurance companies will 

Play in an av future and also, 

You know, how does the existing 

Insurance system as it's 

Currently construed form a 

Barrier to the adoption of 

Emerging autonomous vehicles.

well the lawyers are going to

Have a field day with this 

Because they're going to try to 

Figure out who is responsible 

For this crash?

The driver or is that the 

Machine or is it the 

Infrastructure or just who is 

Really responsible for this 

Crash.

And for a while the attorneys 

Are going to have a field day.

I am an attorney so I am 

Familiar with that.

How the dust is going to set ole

That, I don't know.

I have to think the manufactures

Are going to play a much bigger 

Role in liability because 

They're designing the software.

So the drivers may still be 

Engaged but to the extent you 

Take the drivers out of it, that

Inherently reduces the potential

Liability of the drivers and 

Increases liability on the 

Automakers and infrastructure 

Designer.

So that's going to be a tough 

One to predict.

Stay tuned because that's so 

Transformative it's hard to know

Today what direction that will 

Take.

Sorry not to have anymore c

Cron -- concrete answers but I 

Think that's inherent in the 

Aspect of this being a 

Transformative technology.

so our next question is:  how

Do you feel about an incremental

Approach to implementation?

So basically starting with 

Smaller tasks and then moving on

Up?

Do you think that's a good 

Approach in this newer era?

well that's an interesting 

Question.

There are huge debates.

There's two not so parallel 

Champs.

Some say let's do baby steps.

Some say let's jump all the way 

To the ultimate conclusion.

I don't think anybody is ready 

To jump to the ultimate 

Conclusion because I don't think

Anybody is ready for a 

Driverless car for prime time.

So the question is:  as you go 

Baby step at a time, what the 

Aviation world is finding out is

The closer you get to full 

Automation, the more challenging

The human factors become.

The closer you get to complete 

Driverless situation the more 

Challenging the human factorish 

Yous are.

And one of them is what I 

Mentioned.

Humans are not good monitors of 

Reliable systems.

The more reliable the systems 

Get, the more challenging it is 

To keep the pilots engaged.

You are hearing today about 

Pilots who aren't engaged enough

So they lose their basic skills.

In 447 they didn't have the 

Basic skills how to fly at 

Cruise altitude because they 

Relied on the automation.

There are pros and cons.

I don't think anybody is ready 

For a leap to full automation.

But, on the other hand, as they 

Take the baby steps toward it, 

The closer the steps are to 

Removing the human the more 

Challenging the conditions.

I prefer baby step becausey 

Think anybody is ready for 

Prime time.

The theory is simple.

Most of this is human error.

Nitsa says 96% of the accidents 

Are human era.

To remove the human and you 

Remove the human era.

Well that sounds good but that's

Over simplistic.

It's not going to be that 

Simple.

We will have to do it in 

Measured steps.

And I think the feds need to 

Play a role in every one of 

Those steps as opposed to just 

Letting it happen.

I mean, in terms of creating 

Constraints but I think the feds

Need to be on top of it so they 

Can play the active role they 

Need to play not only in the 

U.S. But in the worldwide 

Community.

could you provide any new 

Lessons learned that would come 

From the investigations into the

Downing of the two boeing 747 

Max plane as they relate to av 

Adoption?

I'm not sure how much these 

Are related to av but I can just

Tell you the big picture view of

The certification process, 

That's what I worked on was the 

Certification process for the 


The big picture view is the 

System for improving airplanes 

Has worked very well for decades

And that's reflected by the 

Exemplary safety record that 

Aviation has of going for almost

Ten years and nine or ten 

Billion passengers without a 

Single passenger fatality.

That's an amazing 

Accomplishment.

So a foundation of that 

Accomplishment is a very good, 

Very safe airplane.

The foundation for a very safe 

Airplane is a very safe system 

For certifying the planes.  What

We found was the system worked 

Well.

But what it needs to do because 

Airplanes are getting ever more 

Complex, as the airplanes get 

More complexed the system needs 

To be revisited because these 

Two tragic crashes show the 

System is still letting things 

Fall between the cracks and they

Need to upgrade the good system 

To make it better.

By the way this is not an 

Faa/boeing issue, this is a 

Worldwide issue for all 

Governments.

I give kudos to the faa for 

Studying this with a worldwide 

Group of certification experts 

From basically every certifying 

Authority in the world except 

Russia to come together, them 

Plus nasa, and I was asked to 

Lead that effort.

It's called the joint 

Authorities technical review.

I was asked to lead that effort.

We came up with recommendations 

To the faa that we gave them 

Last october about how to bring 

The system up to date so it can 

Respond more effect lively to 

Today's reality of complex 

Airplanes.

Arm ape not sure there are 

Lessons learned for the aviation

World except you can say the 

More complex the vehicles become

The more sophisticated the 

Regulation method has to be to 

Respond to that new complexity.

the next question we have is 

Asking:  who do you think should

Participate in decisions of 

Ethics as resulted to automated 

Vehicles?

To what extent should citizen 

Expertise be valued by 

Technologists and wreck 

Regulators.

I think it needs to be led by

The feds but needs strong input 

From academia and the driver 

Community as well.

So I think all of those need to 

Be in this picture.

I'm not sure exactly how that 

Would work and whether there's 

Any mechanism to do that but I 

Think it needs to start with the

Feds and I think academia like 

You guys, because I've seen 

Ethics work going on in many of 

The universities and I think 

They need to be involved with 

It.

And by the way not just u.S. But

World wide.

So, I don't know of any 

Mechanism today to do that and 

Not only get input from academia

But from the driver public.

That's one of the things from 

The driver public is the quality

Of drivers is very -- that's a 

Very big bell curve.

So, trying to sort of get a 

Representative am sample is a 

Challenge.

in speaking with your 

Experience with the ntsb, can 

You talk about these 

Stakeholders that you involved 

When you were going through your

Decision making process or your 

Recommendation process?

I think that would be helpful 

For us to get more context.

sure.

The way the ntsb works is the 

Party system.

When they're investigating a big

Airline crash we have what's 

Called the party system.

Everybody who has a dog in the 

Fight is involved in the 

Investigation.

The airline, the pilot's union, 

The air traffic control, the 

Airports, the engine designers, 

The mechanics, flight 

Attendants, everybody who has a 

Dog in the fight would be 

Parties.

The reason for the parties is 

Because they have technical 

Knowledge that the ntsb wouldn't

Have.

So if we're looking for what's 

This piece of metal in the 

Middle of the field, ntsb would 

Have no idea.

But the person from boeing would

Say I know exactly what that 

Piece is.

Or if we talk about how do the 

Pilots usually behave.

The ntsb wouldn't have any 

Notion of how the pilots would 

Behave but we could ask the 

Pilot's union who is involved in

The investigation for their 

Advice on how that works.

So the ntsb does it with a 

Collaborative approach that 

Includes everybody who has a dog

In the fight for the purposes of

Gathering facts.

Once the factual part is 

Finished then the ntsb moves to 

A non-collaborative approach of 

Developing analysis by itself. 

So the ntsb puts all the facts 

On the website so that everybody

Can see them and we invite all 

The parties to snit their 

Analysis against their facts and

We invite the public.

But the ultimate analysis is 

Done solely by the ntsb.

So we don't get accuse of saying

Boeing unduly influenced that 

One or american airlines or the 

Pilot's or whoever unduly 

Influenced that conclusion and 

Recommendation.

So the analysis is done by the 

Ntsb but for the factual 

Development it's -- it includes 

All the parties because they've 

The technical capability that 

The ntsb doesn't have.

a similar question kind of 

Building on the complexity idea.

So, regulators are currently 

Faced with highly complex 

Systems in passenger vehicles.

Recently several unintended 

Acceleration cases have come 

Before nitsa and were 

Essentially ruled to be 

Unknowable do to the complexity 

Of software.

What ramifications does this 

Have for even more complex av 

Systems?

How will regulators respond in 

The cases of crashes that 

Involve proprietary systems or 

Systems that are unknowable in 

Their complexity?

well thank you for that 

Question.

The good news is that cars are 

Becoming more like airplanes and

They're having more like 

Recorders on the cars that tell 

Us what happen.

So if we wanted to see from the 

Recorder whether anybody was -- 

Today most of the cars, if you 

Start them you have to have your

Foot on the brake for them to 

Start so you don't have that 

Problem.

But the cars would have 

Recorders that would show what 

Was the position of the brake, 

What was the position of the 

Accelerator at the time.

All those kinds of questions.

And might even have voice 

Recording to hear what's going 

On.

The noises in the car.

So the good news is that the 

Cars are more and more having 

This kind of recording 

Capability and it's actually not

For accident investigation 

Purposes but so the manufactures

Can use that feedback.

Because like tesla has 

Continuous feedback from your 

Car to the manufacture and they 

Use that information to their 

Credit to help improve the 

Quality of the car.

So the good news is that as the 

Cars become more technological 

Complex they're going to have 

More recorders to help people 

Like investigators figure out 

What really happened here and 

The totality of circumstances so

We are not guessing as to what 

The cause was.

one thing -- another question

I had was:  we've talked about 

Obviously the government's 

Trying to be prepared ford the 

Advent of this technology and 

How oftentimes regulation or 

Government policy can be 

Reactive in this space.

I'm curious to get your 

Perspective on how government 

Regulation or perhaps any future

Recommendations could help 

Address the equity issues 

Surrounding in safety within 

Transportation and how you see 

That to be changing too.

the equity issues are a big 

Hole that I've seen that is not 

Adequately addressed.

Because I'm looking at to the 

Extent people are going to be 

Beckoning cars, is that 

Something that only rich people 

Can do.

Or scenario anybody can do that.

That depends on so many 

Circumstances.

For example, I'm -- one of the 

Models I'm asking where you 

Beckon the car to take you from 

Where you are to the mass 

Transit then you get mass 

Transit and go the rest of the 

Way.

If that doesn't take you where 

You are going you might have a 

Car at the other end.

That's one scenario.

If it's car all the way to your 

Destination, which I have 

Trouble seeing because when they

Talk about autonomous cars will 

Reduce auto congestion, I don't 

See how many everybody is taking

A vehicle to their destination.

So what I have not seen in any 

Of these models is who is 

Addressing how it's going to 

Affect people of different 

Means.

And I don't see that happening 

Yet.

And I have no clue where that's 

Going to go.

I think it need to be addressed 

But the people with less means 

Don't have as much political 

Clout to make sure they're 

Covered by this.

We'll have to see where it goes.

I'm hoping that people will want

To address the totality of 

Circumstances realizing that if 

These people of less means can 

Get to work more easily we all 

Win.

So it's not a subsidy thing it's

A we're all in it together 

Because it's beneficial to all 

Of us.

I don't have a good answer.

Again I'm sorry not have a 

Complete answer but that's what 

Happens with this very 

Transformative neck neology.

-- technology.

these vehicles will be 

Collecting a lot of location 

Data and will be reliant in some

Instances on that location data 

To be located in space to be 

Properly regulated.

Also we will have -- there are 

Certain models for av that 

Require vehicle to structure 

Connections which provides even 

More reams of location data.

Could you speak to the privacy 

Concerns of location data and 

How this issue is currently 

Being thought about in the 

Regulatory community?

And how it ties in to other 

Federal action on personally 

Identifiable information, data 

Privacy, and location data?

next question please 

[laughter]

That's on the list of issues 

That haven't been dealt with in 

Aviation because aviation 

Doesn't have a privacy issue.

Because the things from the 

Black boxes in aviation 

Blotching to the airline.

So, nobody cab have access to 

Them unless the airline wants 

It.

So, there wasn't that privacy 

Issue.

So that's an example of an issue

That's going to be faced with 

Autonomous vehicles that is not 

Currently faced in aviation.

It will take some privacy 

Protections.

Even today people are worried 

That, you know, with the onstar 

System it will tell where you 

Are.

What if I find out my wife is at

Her boyfriend's house.

There are lots of privacy issues

Associated with that which 

Haven't been -- I haven't seen 

Any serious issues to adequately

Address those but I am assuming 

The political oomph will be 

Enough that privacy protections 

Will be maintained.

It may be they're behind the 

Technology a little bit but 

Eventually I will be surprised 

If they aren't huge privacy 

Protections to prevent misuse of

That information.

so we have a question coming 

From twitter.

So, connected to ought maded 

Vehicles do you have thoughts 

About urban air mobility?

There seems to be a lot of buzz 

About uber elevate and lots of 

Technical and regulatory 

Challenges.  So love to hear 

Your thoughts on this new 

Developing area.

well that's going to be even 

More interesting than what we're

Seeing on the ground.

Because not only do they have 

All the automation issues that 

We're now seeing in cars, but 

They have two additional issues.

First of which is propulsion.

The ones I've seen the geometry 

Is that the vehicle is 

Surrounded by vertical lift 

Fans.

So, let's take the scenario 

Where there's a fan on each 

Corner.

If you lose the fan on one 

Corner you are no longer 

Controllable unless that fan is 

Activated by the other engines.

That's the only single example 

That I know of is the osprey 

Which is a two engine airplane 

And it takes off like this and 

Then the wings go forward and it

Flies horizontally.

Both of those engines power both

Rotors.

So, you still got propulsion for

Both rotors.

If you didn't that would be an 

Instant crash and catastrophic.

Absent some way to maintain that

Control.

And I'll just use the four 

Corner example.

Because I know all you need is 

Three to keep control but they 

Need to be more or lessy 

Metcally spaced around the 

Middle.

If there's three or four on a 

Four corner approach that's not 

Symmetrical.

What are they going to do in the

Case of propulsion failure.

Unlike airplanes which crash 

Over uninhabited ground these 

Vehicles spend 100% of their 

Life over the city.

That means they're going to 

Damage buildings and hurt 

People.

If you remember they used to 

Have helicopter service from the

New york airport to the pan-am 

Building.

One of those crashed -- it just 

Lot a rotor blade and that 

Cascades down and hurt a bunch 

Of people.

And that destroyed the whole 

Service.

If you have one of an urban air 

Mobility vehicle anywhere in the

World that's going to seriously 

Impact.

Number two is now you've got low

Altitude traffic over a bunch of

Vehicles buzzing around and the 

Faa has no way to control low 

Altitude traffic.

So that's going to need a whole 

New infrastructure to control 

The traffic in addition to 

Having collision avoidance 

Systems in the vehicles but it's

Going to take a whole new 

Structuring.

So the bottom line is just like 

I thought ten years ago that 

Predictions of how soon we're 

Going to have driverless 

Vehicles on the street, we're 

Way off.

I think the predictions of urban

Air mobility vehicles in not so 

Long a period of time are 

Unrealistically optimistic.

So, they're going to have huge 

Challenges to face.

The one that I said in uber had 

A big program in d.C. And I sat 

In this taxi which seats eight 

People.

That means it is a pretty big 

Device.

That means if this thing 

Crashes, somebody is going to 

Get hurt.

That's going to be even a bigger

Challenge.

I think the predictions of that 

Are more unrealistically 

Optimistic than they are for 

Cars.

 we have another question 

From twitter.

Since the u.S. Is the largest 

Auto market in the world, why 

Isn't usdot taking broader 

Leadership?

And are there risks involved 

With abdication on the federal 

Level to not set the playing 

Field for av?

Is there possibility for other 

Countries to take the torch and 

Run with it?

there's a big possibility.

And that's is why I'm 

Disappointed to see that nitsa 

Is not taking a more pronounced 

Leadership role in addressing 

This.

Because that leaves a vacuum for

Other countries to take over.

And I think we need to be 

Reeditsers rather than -- 

Leaders rather than followers.

So I am very concerned about not

Seeing nitsa take the lead.

I was the active administrator 

Of nitsa myself 20 years ago.

I don't want to speak too 

Negatively but I am just not 

Seeing them come up to the plate

To take these issues in on 

Automation.

I'm also curious to learn 

About within the ntsb what new 

Technologies you have been using

In terms of monitoring.

Whether you are a proponent of 

Creating innovation of the 

Recommendations you develop.

sure.

And I told people when I took 

Over chairmanship of the ntsb I 

Elide to people the ntsb has to 

Be innovative because the 

Transportation world is 

Innovative.

If we don't continue improving, 

We're going to fall behind.

So I can't tell you how pleased 

I was when I visited the 3m 

Facility in minneapolis and saw 

The management philosophy in the


Has to be innovative.

If we're not we're going to fall

Behind.

Innovative means thinking out of

The box which also means you 

Will make mistakes.

That means I am not going to be 

Hard on you if you make mis

Mistakes.

It tells me you are trying to 

Think out of the box and I am 

Going to praise you and see what

You can learn from those 

Mistakes.

To me, that's essential is if 

The ntsb.

One of the things that we 

Started doing in my tenure was 

We started using drones to help 

Us investigate accidents.

So, now the drones can go into 

Buildings where like we had a 

Collision where the train 

Crashed into the terminal at the

Station and made the building 

Unsuitable for entry yet we 

Could take drones in and see 

What was happen.

Or map the debris field with 

Against on a drone so we -- gps 

On a drone so we know where all 

The debris fell then we can have

A map to use that later to help 

Us figure out the dynamics of 

The crash.

So, we've done a lot of -- we're

Also doing innovation on how to 

Use the information better from 

The flight data recorders and 

The cockpit voice recorder.

So, there's lots of innovation 

Going on at the ntsb.

we're getting a lot of action

On twitter today.

I have another question about --

A couple of the crashes that you

Highlighted.

The tesla crash.

But maybe more appropriately the

Tempe, arizona uber crash 

Involved private vehicles 

Operating on the public 

Right-of-way that were 

Essentially untested or 

Technologies that didn't have 

Third party oversight operate 

The public right-of-way.

What is the current safety net 

To ensure that the public, the 

Traveling public can seek 

Restitution or that local 

Governments aren't held liable 

For public use of -- public 

Operation of these untested 

Technologies?

this is where one of the ways

Where I say the feds need to be 

More engaged.

Because without uniform national

Guidance on that I think we've 

Got a problem.

It's not only for testing on the

Streets but for all these 

Upgrades that keep going into 

Your cars.

As I said, I wonder how 

Adequately was that upgrade 

Tested against the rest of the 

Software that it's entering?

Does that generate unanticipated

Consequences with adding that 

Piece to the software that's 

Already there?

Seems to me the feds need to be 

In on that as well to make sure 

That the software that's added 

Doesn't make the car more 

Dangerous than it was and to me 

That's not a state function 

That's a federal function.

So, we're going to have to 

Figure out a way that before 

This gets broadcast to your car 

It's been tested in a way that's

Adequate to the feds so they 

Will be comfortable with this is

Safe enough to be on the 

Streets.

But I don't see that happening 

Today.

Maybe it will as this evolves 

But I'm not aware of it 

Happening today.

to the point of these -- the 

Data on these crashes, these 

Recent issues, I'm curious to 

See what your response is to the

Pushback from these companies.

From the ubers and eh teslas of 

The world who believe is the 

Future of trainings purples.

How you -- transportation.

How you communicate to the 

Pushback in this area.

if it comes after an accident

That's one thing.

If it's just generic, that's a 

Different scenario.

There's always going to be 

Pushback where the industry 

Thinks we've got this figured 

Out but they don't.


That.

So I think that's going to 

Happen.

Buts that where if the feds 

Don't get into this and play a 

Role, that's not happening in 

The way it needs to.

So I think that needs, again, 

Another reason why the feds need

To get actively involved in this

Arena.

You don't want the states doing 

This individually.

You don't want the manufactures 

Deciding ethical issues.

You need the feds in there for 

Guidance and leadership.

we have another question from

Twitter.

Could you address the carbon--

what's twitter by the way?

Just kidding.

I do know what twitter is.

will automation lead to 

Reductions in carbon emissions?

Or what are the various 

Scenarios for carbon emissions 

And other types of pollutants.

one is the reduced number of 

Cars is going to reduce the 

Pollution.

But I think the bigger answer is

That automation is going to mean

Electric cars for the most part.

Because gas powered cars don't 

Respond to the automation as 

Well as electric cars do.

I've got an electric car so I've

Been there, done that.

Just to take an example.

The purr pulsive efficiency of 

Gasoline powered cars is less 

Than 30%.

Of electric cars it's more than 


So, there's a three times 

Disparity in the pro pulsive 

Efficiency.

So, in terms of how many miles 

You go for this bit of energy 

You can do three times more in 

An pluck car.

So the reason the gas star cars 

Are so far behind because, for 

Example, they generate so much 

Heat.

Not only does the energy from 

The heat come from gasoline but 

Also you need energy from the 

Gasoline to take care of that 

Heat and get rid of it with the 

Fan and the water pump that 

Moves the water through the 

Radiator and the fan that blows 

The heat out.

So it generates more heat and 

It's got to get rid of that.

So that's a big reason why it is

So much more -- the pro pulsive 

Efficiency is so much worse for 

Internal combustible engines.

Also they make a lot of noise.

When I tell a con veteran 

Owner -- corvette owner that my 

Electric car can beat your car 

Across the intersection because 

As you know electric motors have

Maximum torque at zero.

That's why the most powerful 

Tesla goes from 0-60.

I tell my friend my corvette can

Beat your car.

He says yeah but your car 

Doesn't make noise.

So, there are lots of reasons 

Why electric cars are far more 

Efficient than internal 

Combustible engines.

I'm wondering why I'm not 

Hearing a brew ha ha protest 

From the petroleum and car 

Industry when people are 

Predicting 60 to 70% fewer cars.

I don't know why I'm not hearing

It.

To me there's a huge difference 

In pollution from -- I mean, 

What I just told you was a pro 

Pulsive efficiency.

If you want to look at question 

You got to look from beginning 

To end and see what the totality

Of the energy use is from the 

Time you get the fuel out of the

Ground and carried as gas to the

Gas station or carried as 

Electricity to your car in the 

Garage.

So the real answer would have to

Depend on the totality of 

Circumstances.

But I suspect if you look at the

Big picture the electric cars 

Are going to come way ahead on 

That which means less pollution 

And more user-friendly and more 

Sustainable.

could you address the 

Differences in potential 

Emissions impacts of the variety

Of av deployments that you've 

Discussed.

You discussed both a fleet-based

Deployment with on demand 

Service and also the potential 

For a private model.

yeah, again, I think the 

Electric car is going to be way 

Ahead on emissions of the fleet.

And to the extent the car is 

Being used continuously instead 

Of being parked so much of the 

Time, a lot of the inefficiency 

Cars is while they're warming 

Up.

If you have less of that you 

Don't have that concept anymore.

So I think the electric cars are

Going to be the way of the 

Future for autonomous vehicles 

And far better on emissions than

Internal combustion engines in 

The total picture.

so this will be our last 

Question.

Av--

I'm having fun you don't have

To stop.

we can stay after and chat if

You are okay with that.

Av represents a complex solution

To a seemingly simple set of 

Problems.

Are there examples from the 

Aviation industry that have 

Stripped complexity rather than 

Added more complexity that we 

Can takes a lessons learned 

Moving into the future?

I'm not sure I would describe

It as a simple set of problems 

But certainly aviation as it's 

Become more automated has become

More complex which has increased

The challenges.

And that's where the jatr went 

With the recommendations to the 

Faa.

Airplanes are becoming more 

Complex and the system that 

Approves the airplanes needs to 

Be revices and updated to 

Respond to that increasing 

Complexity.

moderator: well let's thank 

Chris one more time for coming.

[applause]

Thank you so much for joining us

Here at the ford school.

my pleasure.

Athreya

and taking so many questions.

You took a lot of questions.

thank you.

My pleasure.

there's a reception out front

Now.

Please join us.

You can ask him even more 

Questions if you'd like.

So, thank you, everyone, for 

Attending.

thank you.