it is that wisconsin book festival that takes place at the downtown madison public public library and october 22. at the state capital another downtown locations. for more information about the book fairs and festivals into what previous festival.

>> this week on "the communicators," the discussion about the guidelines and rules proposed for self-driving cars. our guests, mark rosekind, who is the admin straight tore of the national highway transportation administration. dr. rosekind, what should the public right now know about self-driving cars? >> they're here. the question, not even a year ago, six months, when are they going to arif? they're here. they're on the roads. all of us need to focus to make sure that they are safe as

possible because they offer us tremendous life-saving potential. >> your agency put out guidelines for proposed rules for self-driving cars. what was your focus? >> the policy focuses in four areas. to make it sort of week, the vehicle performance guidance for people that are going manufacture these. if you're going to design, test, deploy, here is a 15-point safety assessment you have to go through to make sure it is safe before it gets on the road. we can talk about this more but it's a total reversal. right now we're trying to look how we build safety into a vehicle beforehand because for 50 years, nhtsa, national highway safety traffic administration is only reactive. you have to wait until somebody gets injured or killed. how do we build in safety? we have a model of state policy we wanted to be specific as we could for the potential to uniform consistent framework of policies across the country.

a patchwork could kill this technology. the last two fell into one area, what tools and authorities does nhtsa need. how do we use current ones to get the safely on the road now? what would the future need to help expedite that safety. >> could you speak a little more about the premarket approval process. that is a big change as you noted. >> and thanks for bringing that up because i think the secretary talked about that a lot because he is really trying to emphasize for 50 years it has all been reactive. so basically we put out a regulation. then we wait to see if you comply with it, or somebody gets hurt or killed, we find out if there is defect, recall, penalize you. it is after the fact. what the secretary focused on, nhtsa, d.o.t., how do we become proactive? how do we get safety in the front? specific to premarket type of approval that is just one of the 12 tools that are identified. we don't ask for it, don't necessarily say we're going to

do it but highlights the future method we might need to get these safely on the street. it is different, will take a while to get there but we need a conversation to explore how we might use new tools for safety. >> joining us in our conversation is dave shepardson, thomson reuters and previous auto reporters as well. >> mark, one of the things nhtsa opted for guidelines as opposed to regulations. was that is function of the fact there are too many unknowns or there was not enough time to go through the regulatory process to create binding rules. >> both of those. let's start with the regulatory side. there are a lot of people for 50 years that you have to regulate and the only way you get your safety. the last big technologies we've had, so that's, rear visibility cameras, talking about electronic stability control we all have, advanced side airbags, six, eight, 10 years, to get rule-makings done.

in six, eight or 10 years with this technology it would changing so is fast it would be outdated before you competent the regulation out. that's why we need new tools. the other area you highlighted there are a lot of unknowns out here. come out with the regulation, what is the performance objective, test data. usually penetration in the market of 15 to 20%, we don't have all that. we have to finish it out. we try to find a safety framework but nuture and support innovation. by having our 15-point list, both come to meet the priority performance items, totally different ways, be innovative as you want, but show us the data. prove you have a safe way. when it comes to regulation we'll pick whoever does it best with data to support it. left's start filling in those unknowns with knowns. >> do you think, what is the

auto industry early reaction? are they receptive to the voluntary requirements? would congress agree to something like preapproval authority for the agency? >> the initial reaction is generally been fairly positive and i take that as ironically people don't think of the government this way but i'm trying to position this as we're trying to be innovative as we do from the government side as we what we try to nuture and support on technology and manufacturing side. it's a little different. some is flipping on its head but we're trying to meet that. so i think the folks who are in this area appreciate the fact we're trying not to jam them in a box with regulation but open it up for safety to see what they can do. and as far as congress goes, those discussions are for the future. and so what we did was identify 12 different tools. each of them probably requires a group to look at what would you do. that happens to be the biggest one everyone is focused on.

the secretary in particular highlights that. it is a example of flipping it on its side from reactive to proactive. there needs to be a lot of discussion before there is any action. >> do you think california will drop its plans to go ahead with its own rules that require a driver in the driver seat and follow in line what nhtsa proposed for the model state guide is. >> for sure you're going to ask california that. i won't speak with them. what i want to highlight the american moat vehicle at administrators. had a lot of actions with states including california. they helped us anticipate what the state issues are. we were interactioning with them with special groups. california is extremely collaborative telling us what their challenges are but also has been waiting to be patient as we come out with ours before they act. pretty soon we'll hear about their next steps.

the indicators they have given us to come out in step with what is our policy. >> dr. rosekind, were there any regulations prior to it your announcement on september 18th. >> thank you for that question because it is so critical. we did our own analysis. we had volpi and part of the transportation administration, how automobiles are handled in current standards. they are not addressed. if you think about it, this is what we wrote, if you manufacture a car that complies with all the current standards and self-certify thaw complied you can put any automation you want on the road. that is why this was so significant. this policy says if you want to do that, here is how you have to address safety before you ever put that on the road. >> why choose self-driving cars for the change in how you approach policy? >> well, and that is another critcan element which is, and this is a number, i'm surprised i didn't start with this, but

i'll do it now, 35,092 number of lives lost in 2015 on our roads. that was 7.2% in crease from 2014. over a decade we you saw 25% drop in lives lost. last year we lost a third of that progress in one year. people will say, keep doing stuff so we can save those lives. it was einstein, keep doing the same thing, expect a different outcome, not going to happen. we need to find new solutions. technology innovations we're looking at in vehicles represents that transformative opportunity to save those lives. everyone has a the potential to be saved. >> given that this is technology, it is in an automobile, but given that it is technology, will there be other government agencies involved in setting rules and regulations about this? >> that's a great question because i think we're, keep trying to highlight, everybody keeps talking about their vision for this space and i think we're

watching it, that future is being created right in front of us. that's a great question, right now nhtsa, d.o.t. we're smack in the middle of it, part of that is secretary fox's vision that we need to play a role. federal motor carrier safety administration, they deal with trucks. they will have a role in this. we're looking potentially spectrum, how that is shared for vehicle to vehicle communication which is rule making we're looking at, connected automation. there will be a lot of government agencies that will be involved. when you start with a vehicles central point, that puts nhtsa, d.o.t. right smack in the middle of it. >> dave shepardson. >> the critics on self-driving cars there is not enough miles. google for example, has driven two million miles. in fact one of the witnesses who testified at the hearing in senate said we need hundreds of millions of miles for testing. it is too early to have these vehicles out on the road. what do you say to that? and do you think more testing needs to be done before we can

be assured these vehicles are safe? >> i hope people start reading the policy and understand how thoughtful it is, and by the way i wasn't at the senate hearing. that was someone else. one of the questions i've been asking what are the new safety metrics we need? we're in transformative place than -- we can't keep counting lives lost. we need new safety metrics. what are those? counting miles traveled, is that really the metric we need. if you read the policy i hope people will see how thoughtful it is, talk about simulation, track, on the record. it really expedites the exemption process to get more on the road for better data. i really think we need more data through different means and different mechanisms. some of them we outline. we'll have more date today in different ways. i still think the bottom line, what are the safety metrics. for example, in california you have to report when you disengage, et cetera. what about a hard braking? what if the hard braking because

somebody stepped in front of the car and they didn't hit the pedestrian. all you see is hard braking bad, but hard braking good. we don't collect data. to collect date that to prove this is saving lives, that is harder to do. that is the metrics we have to figure out beyond miles traveled. >> a company self-certify can put out anything out as long as they meet guidelines. google for example this year it had a crash admitted some responsibility for hitting a us about. tesla had autopilot crash it acknowledged the system did not work as it intended. should these systems be on the road now without nhtsa having some preapproval or more oversight into niece near fully-automated vehicles? >> what you highlighted how fast is this is moves. so in january section take fox announced president's proposal, 3.9 billion for this area and secretary announced within six

months we would come out with this guidance. what is interesting is, how much happened in that time. so if you think about it, it is accelerating literally so fast, trying to get in front of it is a guess for anybody. one of our challenges was to put a stake in the ground that basically says going forward here's what we expect but there is also a section in there that says if you're already on the road we'll give you a little bit of grace period but within this amount of time we want you to fill the same safety requirements now for stuff that is on the road. that is how you get to future but also what is out there now. >> but isn't a question if a company reports to you to meet the 15 guidelines, if they don't, you don't agree, what is nhtsa's recourse at that point? >> this is critical. this is a policy but we have high expectations that people will follow them and one of the things that also came out when

this policy was issued was an enforcement bulletin, when it did, say all is current authorities we have, recall, penalty, all that, that applies to new vehicle technology as well. and so we have a lot of authorities. we could use more that would help us be even more effective. that is in our 12 tools for the future but our current authorities apply to what's going on here and so we have investigations that we can start, we have all kinds of other tools already available for us to try and be sure to safety is built in the front end. >> dr. rosekind who can comment in the 60-day period we're on for the guidelines? >> everybody. that is really important. i hope you send something in. learn bit because we want the public, potential drivers in the future, passengers, there 16-day comment period. there is comment period. nhtsa.gov, you can find it there and anyone can submit any kind of comments.

but i point out in the policy we identify 23 next steps beyond the comment period. we'll have interactive work shops. we'll have an expert review of the policy we'll start working groups to look at some legislative issues. we specifically recommend a commission that the states put together to look at liability insurance issues for example. there is a whole set of next steps that will go for months forward. decides the 60 days there are many other being a activities we expect everyone, manufacturers, media, passengers, anybody should be making comments so we take those into consideration. we made a commitment to annually review and potentially update this policy. >> on your checklist of 15 points for these guidelines, the first one is data recording and sharing. is this a sharing of technology or a sharing of data? >> it is data sharing and i would just say while the response has generally been positive the one caution that people like to raise is what about data sharing? because within confidential business realms people worry

that's especially in this data era, right, there's money there. and so i think what i like to highlight, i'm a recovering nasa scientist, so i come out of aerospace, in the aviation industry where safety is not just a high standard but they have done so well, there is a lot of sharing of safety data. so what we're focused on how do you share safety data. i give you two examples quickly. there are a lot of what people call corner or edge cases. that google, valentine's day, google meets bus, that happens, that may not happen to anyone else. everyone trying to program vehicles should know about the one instance because everyone can learn about that the other part about sharing, talking about all the vehicles on the road that learn all safety information, they all get to share that information. you only know your driving is going to improve because of experience you get on the road. just think of the millions of vehicles that could be adding to the vehicle that is actually

taking you in a driverless car. that is opportunity you don't get with humans right now. >> another one on the checklist is vehicle cybersecurity. we recently visited with raj kumar at carnegie mellon university who is working on driverless cars. here is a little bit what he had to say. >> because they are driven by computers and hundreds of lines of software code, we have to be cognizant of and be cautious about possible threats to the technology. these could include cybersecurity attacks, if i can send information out, i can also receive information back in. those entry and exit lines become potential portals for attack. we have to be aware of attacks that come across the state, the

country, the continent or globally. again besides cybersecurity threats, one also has to be worried about what can be done in a physical context. for example, you can jam gps. you can possibly spoof gps. you can use, for example, even simple laser devices you can buy very cheaply to confuse the light or sensors causing the vehicle to brake itself. so there are cybersecurity threats and also physical attacks that are possible. >> dr. rosekind. >> i think we have to look at both the positive and negatives. let me start with the positives which is one example of how all this technology is over the air updates. you can update software over the air. you can have a brand new car with all kinds of safety features when you wake up in the morning, remarkable. another one, say there is a recall or defect has to be

fixed. everyone wakes up there is 100% recall completion in the morning when everyone wakes up to their car. that is remarkable. what is being pointed out that represents a portal or vulnerability, somebody, any bad actor, concerned what cybersecurity risk i can get to, that opens up. why nhtsa shortly come out with cybersecurity best practice. we identify a whole list of them for that particular point in the vehicle guidance. there is the auto isac that the industry started that we pushed them to begin. this already exists in energy and infrastructure. they already have systems and report and share data. they already share data about these. it is acknowledged. the challenge, this moves so quickly, how do we make sure the best practices are in place to protect people. >> there is a lot of sort of two schools of thought here. that self-driving cars will come faster than we thought. others say there are too many complexities, companies like

toyota say, hey, it is probably a little further away. in terms of what you come down where we see level four, level five vehicles and deployment? more in the 2020 range or further, 2025 and beyond? >> let me go at that by saying i think one of the challenges everyone has is actually trying to segment these a little bit more clearly than the way we talk about them. so you talk about levels. only three of us know what you're talking about. everyone else is like what is that? there is full self-driving. where we expect no input from the humans. the car is monitoring the environment. then there is other levels where the human is still responsible for monitoring the environment or car or some piece of it. the reason i say that i think fully self-driving in all environments, that's probably 2020 beyond. i think we're looking a little bit longer than that. but it's on the road already. so i think the question more is, we may get specialized version of these.

so one of our 15 has to do with operational design demand. what does that mean? where should the car operate? if you have a pizza delivery only during the day on these three blocks, that is very specific application. that could be, quote, self-driving but it is a very specific place with low risk, a bunch of different things. all the way up to the self-driving or the, i think it depends what you're looking at. these things are already on the street. they are going to happen real quickly. full self-driving in all environments i think that is 2020 and beyond before we get there. >> one of the things the guidelines say maybe intuitively that the self-driving car should follow the law. google, for example, as it reports these crashes is getting rear-ended quite a bit because driving very safely. following to law. coming to complete stops at stops signs. should in the future self-driving cars be allowed to drive more like human beings do, going beyond speed limit, not coming to full stops at stop signs and how do we have environment with human drivers

which are less than why thal and self-driving cars following letter of the law? >> i love this points out how complicated this is, right? besides the 35 you thousand 092 lives lost, 94% are human close or error -- 35,092 lives. the irony of that, right? in the policy we address talk about literally down to city laws, there are times you cross a double yellow line. there are emergency vehicle on the side or pedestrians, there have to be times, where you quote, break the law. these are things that these vehicles have to be programmed for as well. what we list in there are a number of scenarios. you tell us others and how your vehicle has been programmed to deal with those specifically. >> dr. rosekind, is self-driving and automated cars is connected cars?

are connected cars a step in that direction or is that a separate issue? >> at the department of transportation and nhtsa we see them as connected and in fact again i give the secretary fox his vision through a lot of this. a lot of people like to say it is one or the other. let's connect the vehicles. what that means your car is talking to another car or the infrastructure. why that is relevant, cars can start seeing around corners. they can see across the intersection. somebody did a hard braking before they got to stoplight. that is things individual person can not do. or others say self-driving fully independent. connected automation where you put them together, gives us a system safer than either one of those separately. so at d.o.t., again the policy just came out. at same time we are pushing we hope soon to push out a rule on vehicle to vehicle communication. the combination of those offers greatest safety for all of us.

>> how, what is nuts is -- nhtsa going to be about the ethical dilemmas that they face? if pedestrian walks in front of self-driving car, should it avoid the pedestrian or protect the occupants of the vehicle? what is your role in that issue? >> that is a great question we decided to highlight because no one has an answer. so i would like for us to get credit not stepping in when there isn't, trying to provide an answer when there isn't one. i'm saying that because we believe that one of the biggest, most important things going forward is consumers, drivers, passengers, accepting these vehicles. those kind of cybersecurity, if my car can get hacked and i'm in danger, i'm not getting in it, right? what about the ethics? by the way that is really nice ethical dilemma. is it going to protect me or -- these things are complicated. people started raising questions but there is not a lot of information. in the policy we identify that as one of the concerns. we specifically talk about how

companies need to address those and be explicit what they provide us to know what issues they address and how they resolve them. as i was talking about innovation. people think that is technology. this needs innovation too. how will we handle those ethical kind of dilemmas in this technology revolution that is going on? we will look to see what people are bringing forward to us and decidesdecide which ones deliver on safety for the public. those are ones we support in the future. >> dr. rosekind, will we enter into a period much like 100 years ago where horse and buggies lived on the same streets as horseless carriages? >> these are, i'm not blowing smoke here, that is the term but these are great questions because we don't talk about that enough? i actually think there are two transitions that don't get discussed. one is in the vehicle. some people are focused on that is your kind of level three, you're human. you still have to be knowing what is going on. if there is something going an

you might have to take over the driving. there is transition from automation to human. a lot of questions how that will go. the other transition what you highlighted. average age of cars, 11 1/2 years. if you had a perfect self-driving car tomorrow it can take 20 or 30 years to have the new technology penetrate the fleet. full turnover, 20 or 30 years. if you have a perfect one tomorrow, it will still take 20 or 30 years. of course, dave says, i really like having top down on pacific coast highway in california on beautiful -- don't take my driver's license away. so what you just described. i think for one, two, plus decades at least, we're going to have a mix of human drivers and automation on the sail roads. and we just said 94% of the those crashes are the humans. that's what we're trying to eliminate as we can. that is translation of mixed fleet we have no data on i think

is future for next several decades. >> after those decades do you see when self-driving cars become so much more reliable pressure on vehicles will not be there human beings fall asleep, who drink and drive, who don't pay attention to driving, can you imagine 20, 30, 40 years from now the government or state say, human beings, you're not careful enough to drive? >> again this is why i love these questions. you see the irony of these right? you talk about will we get into autonomous. if we get to zero why would we let risk of a human take over the wheel right? i don't know that we'll ever get there. that is a question for 30 years from now when there is really data to show how far have we really come. i hate to say it, but probably my kids figuring out are we ever going to it had the point where

that question becomes a reality that gets answered. >> why is this coming at us at a whoosh rate now? >> here is the personal part. i'm third generation san francisco from silicon valley. so i say that because for me, you know, technology is like, what phone version do you have? and why are we talking phones? we should be talking about all the rest of it. i literally stood in line for the first iphone, okay, with my son, you know. i say that because the whoosh we're getting is the way technology happens. and that is what i'm saying. silicon valley, what is the surprise here, people? there is some kid in a garage or their room somewhere thinking up next startup that is going to save more lives. that is the excitement of it. i think part of what has been so interesting to see how there has been so much focus on the road safety side. and so for all the technology that has been the whoosh, think about it, we have phones in homes and internet things and all this stuff. what is fascinating now is the focus on the vehicle because

people now realize safety, mobility, as we get older or you're disabled you know, sustainability for the whole planet. so the secretary's talked about, think about expansions and all other things without pouring another truck of concrete because of ride sharing, right? when you look at those, it is like a no-brainer we need to be going at this as strongly as we can, safely but we can not miss the opportunity presented here. i think it is fascinating how it is picked up by different segments and -- i actually think the technology will not be the issue in the end. again from silicon valley, they will figure stuff out. you got a question? they will figure it out. the issue you're bringing up about the human, will they accept it? what about the ethical issues? would you ever want to give up your driver's license? that will be fascinating part to watch in coming years. >> mark rosekind, administrator of the national highway traffic safety administration.

dave shepardson covers technology for thomson reuters. and over the next two weeks "the communicators" will travel to both pittsburgh and to ann arbor. we visit carnegie mellon university. take you for a ride and show you the technology in autonomous vehicle and we look at connected cars at m-city, part of the university of michigan in ann arbor. thanks for watching. >> admiral john richardson talks about maritime security and naval operations later this morning at the center for strategic and international studies. we'll take you there live in about an hour 1/2. later the association of the u.s. army holds their annual meeting in washington. several senior officials speak to media throughout the day.