Skip to main content

tv   Washington Journal Michael Farren  CSPAN  March 31, 2018 9:32am-10:01am EDT

9:32 am
today, we continue to bring you unfiltered coverage of congress, the white house, the supreme court, and public policy events in washington, d.c. and around the country. c-span has brought to you by your cable or satellite provider. "washington journal" continues. host: our guest, michael farren, a research federal at the mercator center at george mason university. mercator's is the and what is its worldview. guest: it is a research center whose goal is to be the bridge between academic ideas and actual public policy, to help public policy reflect -- to get better over time and help policymakers understand the core pieces of research. host: the topic is the future of our thomas -- common us -- autonomous vehicles.
9:33 am
let's go to tempe, arizona, where the recent news of a fatal accident of an autonomous vehicle, a person was crossing the road. here is a look. explain exactly what happened. we saw the outside view of the pedestrian walking a bike across the street and the internal shot of the car. what happened? guest: it looks like it was a failure of the hardware or the software, and the human backup driver as well. the system should have picked up the person crossing the street just as an object in the street if nothing else. there was multiple system failures.
9:34 am
host: one of the headlines that comes out of it is -- what has been the fallout? guest: uber has suspended its own av testing, and other companies using the same software has suspended their testing as well. companies like way mel have doubled down and are moving forward -- waymo have doubled down and are moving forward. host: this particular car had a driver in it. what does that mean to the concept of an autonomous car? guest: the driver is a backup system because autonomous -- fortunately, that did not happen. part of the question about this is, could this have occurred even if there was not autonomous technology involved? host: how else is the industry
9:35 am
responding? where they'll be -- will there be any kind of chilling effect? guest: i think they're well in terms of firms will take safety much more carefully, and that is a good thing. if there is a silver lining to this tragic incident, it is that we will probably get more cautious and safer in the future. host: phone numbers on the bottom of the screen for our guest, michael farren, talking about the future of autonomous vehicles. we asked him on because of the first fatal accident with that uber vehicle. the numbers are on the screen. you are welcome to tweet as well. role of big governments at this point when it comes to regulation? guest: the role probably could be best described as ensuring there is responsibility for autonomous vehicles and the companies that operate them, in
9:36 am
maintaining safety, but not to specify any particular regulations in terms of you should use this technology or this process in doing the right thing. we need an environment of what might colleague calls her permission-less information in an environment where companies are held to liability standards for the problems they cause. host: what is the promise of autonomous vehicles? where does the industry hope to go? what does it bring to the country? the way that i look at autonomous vehicles is that they are going to make us safer, wealthier, and they are going to give us a lot of quality of living, increase that and give that back to us. we have currently about 37,000 people dying every year on the roadways. 6000 of those are pedestrians.
9:37 am
about 100, that is people a day and 16 pedestrians a day. autonomous vehicles do not get drowsy, distracted, or drunk. we have this opportunity to, once we get the technology correct, vastly reduce the number of people dying on the roads in the u.s. and around the world, where you have about 1.2 5 million people dying every year. host: our guest is licensed as a professional engineer with a masters degree in transportation engineering, a bachelors in and a phd inring applied economics from ohio state university, michael farren, research fellow at the mercada center. we go to barbara from gladwin, pennsylvania. caller: i believe one of these accidents occurred because the person was jaywalking.
9:38 am
you still have to obey the traffic laws, whether you have an automatic gizmo or not. someday, that may change and it may be able to think, but right now, obey the traffic laws. host: anything else? caller: no. host: any reaction? guest: about the governors highway safety association reported in research, they found about three quarters of people struck, pedestrians struck and killed are walking at night and are jaywalking. you are right. it is a fundamental problem, and that is one of the great things about autonomous vehicles, that they can see in the dark. lidar can see in the dark much better than humans can. obviously, there was a failure in this situation in arizona, but this is the kind of thing autonomous vehicles can help out with. host: some tweets by the governor of arizona.
9:39 am
two federal investigations will help. these are independent and led by experts. i want to see the facts. it has all been about public safety. can this reduce fatalities? human case fatalities keep going up. arizona has allowed this technology to test and flourish, but public safety is first and has always been our focus and will remain our focus. how many states are allowing testing? guest: before -- between five and 10 states. arizona is a good one to do it. they are the sixth worst state in the nation for pedestrian fatalities. the good part of phoenix, the phoenix area, it developed late as a city so it has developed more with the car in mind. that also means it is less friendly to pedestrians, very long blocks with long distances
9:40 am
between pedestrian crosswalks. some cities around phoenix like mesa are installing special software and technology to allow pedestrians to cross midblock. that is another thing we can do to keep them safer. host: are you concerned about maybe a patchwork of different states' rules and regulations not being effective as opposed to something federal? caller: i think there's a lot of opportunity for states to come together and work on this kind of thing. it does not always need federal input and federal regulation. interstate compacts are something that are already used for driver's licenses, to coordinate drivers licenses across states. you could do the exact same thing with driverless vehicle regulations. host: let's hear from michael from herndon, virginia. caller: legally, who is responsible for this person
9:41 am
being struck by this car? uber or the extra person that was behind the wheel? questionat is a great and it is something honestly that is going to be determined more by the courts going forward. uber has already settled with the victim's family, so they seem to be taking responsibility. thatis kind of something the law will be working on going forward, which the court system has a good job of doing as technologies and new things emerge in our society. we can model our way through, rather than trying to say, this is how we will deal with it from the get-go. we need to figure out the way to allow us to involve forward -- evolve forward rather than say how we are going to evolve. host: from west milford, new jersey, dale. caller: i have been in software
9:42 am
development for like 30 years. i was a little bit amused when i heard you say the accident was due to either hardware or software, because in reality, anything that involves computers like this is going to be either hardware or software. the environment itself, the radio environment is going to be constant. i think what this accident did is provided a key point to some of the ethical dilemmas that this presents. one of my coworkers this week was saying, imagine, for instance, you have two vehicles traveling on a two lane road and all of a sudden an obstacle appears in one of the lanes. one vehicle has one person, the other vehicle has five, and only one of them can safely pass through. the other has to go off the road to certain death.
9:43 am
clearly, you would have a computer algorithm that would make that decision. the obvious answer of course is that you would sacrifice the one person to save the five. my answer was, who is richer? was the person in the single car the one who wrote the code? if he wrote the code, he has probably put in a special key identifier so they will be exempt. there is a lot of moral dilemmas here and artificial intelligence, people might say that is the way to go, but that becomes even more confusing because then things which we interpret, a computer is not going to. it will make a binary choice, yes or no. guest: this is an interesting question that has come up a lot in debates about this, and i think that a good way to address that to recognize in fact these kinds of situations will be vanishingly rare.
9:44 am
in this situation the caller described, two cars moving toward each other on a two lane road and an object appears out of nowhere, a human will react much lower than a computer driver. the reaction, the perception time, reaction time, and stop -- time to stop the vehicle will be less. we deal withs today in terms of what do i do will become much more rare. while i think this is an interesting theoretical question , it doesn't matter as much in terms of practical terms for what we are going to do in the future. host: you have written that self driving cars are really here and just might save your life. how good is the technology now and how much better can it get? guest: it depends upon the company. obviously, the technology with uber just causing the pedestrian death, there is a problem there. implementation has caused a problem. on the other hand, tesla's
9:45 am
technology has proved to be pretty good in the fact that it has so far shown to be about over three times safer than driving a normal car. there is different pieces of it that need to be fixed. there was an accident in california recently where another driver lost his life, and it may have been due to the technology essentially following a line in the roadway that was not a lane line, but a joint between two pieces of the roadway, between asphalt and concrete. there has been good reporting from abc seven in california out of this. in the future, we are going to world, justsafer like we have a much safer world because of automatic backup cameras today that are now being required automatically in cars, or will be in the near future. we should allow the small problems that we are seeing today -- and don't get me wrong,
9:46 am
these families will be missing their loved ones for the rest of their lives -- but we are going to save tens of thousands and millions of lives in the future by getting this technology right. host: i wanted to get your opinion on this. if people wanted to ban planes after the wright brothers crashed, the world would be a much worse off place, but there is a difference joined the wright brothers testing which took place in empty fields and beaches, and autonomous car's on public streets. as world changing as this technology might be, let's not forget these are public -- private companies. guest: absolutely. that is a key question of what to do. i would compare this rather than to the wright brothers, to the crash in paris. many have seen this picture of the locomotive sticking outside
9:47 am
of the building. that was a failure of both software and hardware. the engineers were driving too fast into the station and relying on the new brakes, the air brakes of the train to stop it and it did not. there was a woman killed in that situation as well, but we did not stop trains as a result. it is a horrible situation, we need to get it more correct, that the future is much brighter for the kind of testing we are doing. the kind of things we never see are the people that are saved by the technology. we do not know how many pedestrians have not been struck as opposed to a human driver who would have missed it. -- asas we take a look at we go to a caller in chandler, arizona, here is a picture. good morning. caller: good morning.
9:48 am
i would like to say that i live in chandler. i do see all types of waymo cars being tested. i kind of steer clear for now. we have a grid. everything is a grid. to bear cars were taken creek road in santa cruz, i don't think they would stand a chance. i would like to see a lot more testing before they are turned loose. i appreciate your comment. host: thank you. guest: yes, definitely. phoenix is probably one of the absolute best places to test cars because of the great network, because it is such a car centric city, and it does not have the kind of weather problems the rest of the country experiences. i was talking with mike uber driver on the way here this morning. he said, would you rather have an uber driver or do an autonomous vehicle?
9:49 am
given that arlington, virginia and d.c. are laid out in a different fashion than phoenix, i said, i would rather have you, but in the future i think i would rather have an autonomous vehicle because they will that -- they will be that much better. host: will these vehicles obey limits or willed passengers be able to override the speed? guest: i don't know. that is one of those questions were somebody else will figure that out. host: james in atlanta, you are on with professor michael farren. do is: what i want to just continue to test and expand it. drive --r, 2000 people die in cars that have drivers in georgia. hopefully it will get better. 1, 2 people every day, maybe three people a day are dying on the highway in georgia. then, it isw and
9:50 am
not the greatest comparison, but we have just got to keep expanding this technology. thank you. guest: i completely agree with that. in fact, my sixth grade teacher actually died in a car accident and it was really traumatic at that point noticing how her husband -- they were both very young -- losing the love of his life, and he was absolutely inconsolable. we need to remember that this kind of technology will prevent that kind of tragic loss in the future, and we need to get it right. the companies that cause problems need to be held responsible. the future is much brighter because of what we are developing. host: just under 10 minutes left, catskills, new york is up next, ken. .aller: thank you both i am curious about the future of employment, frankly.
9:51 am
i am wondering if you could comment on that. i think it is a definite upside to this kind of software, but i am stumped to imagine what the world will look like in 20 or 30 years. guest: to address that, i would say that yes, there are a lot of people employed in driving related jobs. but the way that an economist would look at a job is it is a cost of producing something rather than a value. a normal social analysis of a job is that it is a good thing, and that is understandable, but the idea is that once we have this technology to allow us to essentially provide the driving service without the person involved to free up that person to produce more, to build more things, create more things. we will see more economic expansion and we have essentially more human resources to work with.
9:52 am
it is going to be an adjustment factor. changing over from horses to cars was also an adjustment factor. we had entire industries supporting a horse and carriage industry and we had to shift over to cars, so we are going to go through another major shift but it will be a good thing. host: is there an environmental plus? guest: certainly. you will get better gas mileage out of eventually wider autonomous vehicle adoption. you will have less smog and soot as well. essentially, in addition to the safety, you are going to have a better living environment. host: we have touched on uber and others. who else is active now in this part of the business, the autonomous car business? guest: it is actually hard to name a company that is not getting involved in the autonomous car business.
9:53 am
is, all of the european car companies, so the idea is that it is going to be the way of the future. we need to allow that companies can experiment with this, but hold them accountable so that they understand they are not just simply experimenting in the public space. the value of being in the public space, nothing can compare to it. nothing that we can construct out on a test site can possibly compare to the data that we gather in the public space. one additional value that autonomous vehicles offer is essentially you and i, when we learned to drive, we probably made a lot of the same mistakes over and over. computers do not necessarily make those same mistakes over and over, so it is not just a matter of we will get safer than we are but we will get incrementally safer and less risky overtime as computers no more and more. host: as an engineer, how do
9:54 am
autonomous car's deal with weather, mountainous terrain where computers may not work? that is a long question that you could spend three hours on, but how is that evolving? guest: you would have to bring in one of my colleagues from ohio state who was working on lidar in the yucatan rain forest and using it there. they do have problems with rain and whether another increment lementions -- imc;e,em conditions. eventually, we will end of what the system that is able to deal with that kind of weather. host: here is an interesting question from sandy beach. should intoxicated people the prohibited from getting behind the wheel of their autonomous vehicle? guest: that is a good question. intoxicated people should
9:55 am
certainly be allowed to be driven by an autonomous vehicle in the same way that we use taxis and uber to get home from bars now, people do. whether an autonomous vehicle is a complete tear for e foricated -- cur intoxicated driving, only if the person is able to take control of the vehicle. host: richard in bethel, new york. caller: the question i have has to do with safety. since this is a level e problems and someone can die, how much do you have a redundancy and what redundancy was on this vehicle that caused this wreck? guest: i don'ta fix of the actual uber design, and i imagine that is kind of an industrial trade secret. r, radar,ey have lida and cameras trying to back each other up, but different
9:56 am
companies have different numbers of sensors. the uber volvo suv that was involved in the accident only involved,dar sensor and more cameras and radars as cars whereas the waymo have six or seven lidar sensors in addition to radar and cameras. not more is is automatically better, as may be the one that uber was using was more powerful than the six or using computed as for the backup systems, there are backup systems and in this case the human backup driver system failed. host: we have one more call coming in. speak to us briefly about congress. has congress touched this issue in any significant way and what might it be doing in the years ahead? guest: congress has definitely
9:57 am
looked at the issue and different members have discussed it, but congress is not moving in any major way right now. it is working with the highway regulatory agencies and companies to figure out how to do this new technology. ithost: last call, jerry from et brunswick, new jersey. caller: good morning. i just turned it on, so i apologize in advance if this question has been asked before. there was a recent incident in san francisco in which a passenger in an autonomous and my was ticketed, question really relates to whether there has been any consensus in either the insurance industry or in law enforcement regarding the liability of people in an autonomous vehicle. thank you. guest: that is interesting. i am not familiar with the story of a passenger being ticketed. that seems odd to me.
9:58 am
it kind of seems like ticketing a passenger in a taxi. if by passenger you mean driver that was the emergency backup driver, that kind of makes sense because essentially that driver is the equivalent of a taxi driver. that is about all i can say. host: our guest has been michael farren, part of george mason university. thank you for your time, expertise, and insight. guest: thank you, sir. host: thank you for calling in. we will be back tomorrow morning at 7:00. guests include a look at politics in the media. we will also take a look back at the year 1968 and the issue of race relations. 1968, --es is called "
9:59 am
austin history professor peniel joseph. enjoy your saturday. we will see you back here tomorrow. have a great day. ♪ announcer: next, c-span
10:00 am
testimony on a former cambridge analytica director. later, privacy concerns that have resurfaced after the cambridge analytica revelations followed by a panel discussion on how consumers, corporations, and government purged surveillance and privacy. a former cambridge analytica research director testified before a british parliamentary committee on allegations that the u.k. data from exploded personal data of millions of facebook users to influence u.s. and british elections and may have swayed the u.k. brexit boat . this is 3.5 hours. -- vote. this is 3.5 hours. >> welcome to our inquiry into disinformation and fake news, i

18 Views

info Stream Only

Uploaded by TV Archive on