History

 Until Henry Ford developed the Model T, automobiles operating on internal combustion engines were few and far between. Karl Benz built his first automobile in 1885, and production began in 1888, after his wife proved that the horseless carriage was suitable for daily use by taking a long distance trip from Mannheim to Pforzheim and back. The Model T was considered the first affordable automobile, and it opened travel for the average American. In 1930 the car radio was born. From the beginning, people wanted entertainment in their vehicles.

 In the early 1950s a rather forward thinking person, Dr. Hunter Shelden, proposed retractable seat belts, recessed steering wheels, reinforced roofs, roll bars, door locks, and passive restraints such as air bags. Congress passed legislation in 1959 requiring all vehicles to have certain safety standards.

 From there numerous options for the comfort and enjoyment of the driver and passenger were developed, including air conditioning and heat, power windows, power seats, cruise control, heated seats, and other options. Add to these options anti-lock brakes, air bags, safety bumpers, and stability control, and you have a modern automobile.

 While this automobile is much safer than its predecessors, there are still thousands of fatalities every year. In 2014 32,675 people were killed in roadway crashes; estimates for 2015 are 38,300, the highest since 2008. Injuries were estimated at 2.3 million, making vehicle safety a constant concern. Technology, which is rapidly making changes to the way vehicles operate, has great potential for increasing safety as well.

 Safety

 As mentioned earlier, thousands of people are killed annually on American roads, and millions are injured. Alcohol impaired driving accounted for 31 percent of those fatalities. Distracted driving accounts for 10 percent of fatal crashes, and 17 percent of injury crashes are related to distracted driving. Other causes of crashes involve user error and careless activity such as following too closely, speeding, aggressive driving, or over-compensation. In 90 percent of crashes, the primary factor the behind crashes are human errors. Some experts feel that the statement that 90 percent of accidents are caused by human error is an overstatement; human error is used as a cause when no other cause is obvious.

 Regardless, a vehicle that drives itself, an autonomous vehicle (AV), would compensate for these human errors and could make the roads much safer. Volvo has a goal that no one be killed or seriously injured in any of its vehicles by 2020, only four years away.

 However, the transition between autonomous mode and driver mode can be tricky—a driver crashed a test car during transition in San Francisco on January 8, 2016. Google reports that it can take up to seventeen seconds for a driver to respond to alerts to take over, and there may be some confusion as to who is in control of the vehicle. A danger of AVs is that by encouraging passengers to text, read, or even sleep, how can that driver be engaged in a split second to take over driving in an emergency? The question of handoff is a huge issue in the development and operation of autonomous technology. In a situation like the one in January, the question will be was the vehicle at fault, or did the driver interfere with the auto's operation and cause an accident that would not have happened had the vehicle been left alone? Accident reconstruction will be critical in sorting out such accidents. Redundancies in safety systems will be critical to ensure foolproof operation.

 Some automakers are having the vehicles prompt the driver to take over after a brief period. Nissan admits that even their most advanced vehicles will not be able to act autonomously in every situation, for example snow, heavy rain, and some nighttime driving.

 After six years and two million miles of testing, on February 23, 2016, a Google autonomous vehicle struck a municipal bus in a minor crash. The vehicle was trying to get around some obstacle in its lane, and moved into a lane to the left striking the side of a bus. The test driver thought the bus was going to slow down or stop to allow the Google vehicle to merge into traffic which it did not. Google admits to some responsibility and has adjusted its software. There are obviously issues to still be worked out. Merging is a common day occurrence; since the autonomous vehicle didn't yield then the driver should have taken over, but he did not. Autonomous vehicles do not have the intuition that a human driver has, at least not yet.

 Another safety issue is that while the AVs obey all traffic laws, as other drivers on the roads do not, the AVs can actually cause others to crash because they are so obedient to the letter of the law. The accidents AVs are involved in are often rear-end collisions when another vehicle hits the AV because the human drivers are aggressive or not used to vehicles obeying speed limits and all rules of the road. If an AV is programmed to wait at an intersection until all vehicles come to a full stop, what about when the other vehicles roll the stop, or don't stop long enough for the AV to register the stop? It could cause more traffic issues than it solves. This causes a dilemma for the engineers; do they program the vehicles to bend the rules on occasion, and if so on what occasions? The vehicles also react faster than human drivers do, so will stop when they see a pedestrian sooner than a human driver would.

 A March 2016 review of Federal Motor Vehicle Safety Standards (FMVSS) shows that there are few barriers for autonomous vehicles to comply with existing regulations as long as the vehicle design is not significantly modified. In order for vehicles to be sold in the American market, the manufacturer must certify that the vehicle meets specified performance requirements laid out in the FMVSS. The standards focus on crash avoidance, crash worthiness, and post-crash survivability. The review showed that while many standards refer to drivers, these standards do not preclude certifying an autonomous vehicle as road worthy as long as it allows a human to operate the vehicle with wheels and pedals. The definition of driver is "the occupant of the motor vehicle seated immediately behind the steering control system" there is nothing that says the occupant has to actually be driving the vehicle.

 A recent study by the RAND Corporation states that driverless cars cannot be declared safe until much more testing has been done. The report concluded that the vehicles would have to be driven hundreds of billions of miles in order to prove that they are more reliable than human drivers. Americans drive nearly three trillion miles every year; making the two million testing miles driven seem insignificant.

 Technology

 Volvo intends to have AVs available by 2020, and Nissan states that it will have ten new autonomous vehicles in the next four years. Elon Musk of Tesla states that the Autopilot feature in the new Tesla is better now than a human driver, and that within one to two years it will be possible to summon a Tesla from the opposite side of the country. While initial costs may be prohibitive, it still may be possible for AVs to be available on the mass market by 2022 or 2025. General Motors has invested $500 million in Lyft, one of the ride-sharing companies with the hopes of creating an on-demand network for autonomous vehicles. The idea is that the AVs will transport passengers to their desired destinations while providing Wi-Fi or television.

 The National Highway Transportation Safety Administration (NHTSA) announced in March 2016 that they have received commitment from several manufacturers to meet the goal of making automatic emergency braking standard on their new vehicles by 2022, and on heavier trucks and SUVs by 2025. The use of this advanced technology as a standard feature in many new cars indicates how significant and valuable these technologies are to the safety of people as a whole when on the roadways.

 One of Google's goals is to have vehicles so capable of autonomous driving that they can no longer have steering wheels or control pedals. Such a vehicle is controlled by the self-driving system, which should be considered to be the driver in event of an accident. Upon reviewing such a proposition the National Highway Transportation Safety Administration stated that the SDS would be considered the driver, since it had control of the vehicle in event of an accident. This is advanced technology, and at the moment California's proposed regulations for autonomous vehicles requires a licensed driver be present at the wheel, making an actual steering wheel mandatory. While many embrace autonomous vehicles many are still wary; vehicles without steering wheels and pedals are likely going to be the second stage vehicles, once autonomous vehicles are regularly on the road and regulations have been sorted out. A vehicle without a way for a human to intervene must be accurate 100 percent of the time.

 Volvo plans to provide Gothenburg Sweden with 100 vehicles to actual customers to use in certain areas in 2017. Fully automated parking will be tested as well, where the driver can leave the vehicle and it will find a parking spot and park itself. Volvo is also sending 100 autonomous vehicles to China for testing in certain cities once the Swedish test has been completed.

 The University of Michigan Transportation Research Institute is currently running a pilot test using 3,000 vehicles driven by Ann Arbor residents. The vehicles are not driving themselves, but are collecting data by interacting with each other and sensors that are posted at busy intersections, around sharp curves, and on local freeways. The goal is to expand the capability of the sensors so that they can have decision-making power. The ultimate goal is to have a fleet of connected and AVs on public roads by 2021. Autonomous trucks could move products twenty-four hours a day without having a driver that needed to rest for a certain length of time. Elderly drivers could get around safely without having to worry about vision loss, hearing loss, reaction time and other factors of aging that impact their driving skills.

 There are many steps, however, on the way to fully autonomous vehicles. The National Highway Traffic Safety Administration (NHTSA) had revised their initial identification of autonomous vehicles from 0-4 to now driver assistance, partial automation, high automation, full automation.

Driver assistance is where the vehicle never takes control, but gives information essential or useful while driving to the driver, such as clear and concise warnings when the situation begins to become critical. This includes lane departure warnings, or collision warnings but the driver remains in control, the vehicle does not stop itself. This is your basic vehicle with no system that can control the vehicle on its own. The vehicle may issue audible or visual warnings to the driver, but it is the driver's responsibility to make corrections as needed.

 Partial automation consists of systems that autonomously intervene when the driver fails to act despite warnings; automatic emergency braking, emergency lane assist. Here the vehicle steps in and takes control when it senses an accident is about to occur.

 High automation is where the driver can disengage physically and have both hands off the wheel and feet off the pedals. The driver must, however, be ready to take over at a moment's notice. This is where primary controls work in unison to relieve the driver of control–for example adaptive cruise control and lane centering combined together.

 Full automation is a truly autonomous vehicle where the systems do not need a driver and all passengers can be engaged in other activities. The vehicle can even maneuver without a passenger, something that make it possible for autonomous fleets of taxis to exist. The ability of a disabled person to call for a car to come pick them up would be a significant help. An empty vehicle will arrive at the person's home for the person to enter and ride in to the specific destination.

 While various manufacturers are making statements that by a certain date in the future they will have a fully autonomous vehicle available, that day is not yet here. However, there are several semi-autonomous technologies that are available in 2016 and newer models that are worth noting. Two years ago the technology tended to be on more expensive models of vehicles, and not every model has such options. However every year more and more vehicles are enabled with such capabilities, and on less expensive models as well.

 The first such technology is a forward collision alert system with crash imminent braking. This system warns the driver of an impending front end collision and applies the brakes to try to avoid or at least lessen the impact of such a collision. The system scans the vehicles ahead to advise of potential hazards.

Adaptive cruise control allows the driver to not only set a cruising speed, but a distance the vehicle is to keep between it and the vehicle in front of it. The vehicle adjusts speed accordingly to maintain the distance.

 Lane departure warnings give audible warnings if the vehicle heads into another lane without the turn signal being used. Some vehicles take this a step further and actually brake the wheels on that side of the vehicle to steer the vehicle back into the proper lane.

 Side blind zone alert alerts the driver when there is something or someone in the driver's blind spot.

 Rear cross-traffic alert warns the driver when something crosses its path in the rear when the vehicle is in reverse. Some vehicles have rear cameras as well as the alerts to show the driver objects or vehicles in the way. Rear park assist is an option that issues an audible warning if the vehicle is within a certain distance of another object; this allows the driver to make adjustments before it is too late.

 The following is a sampling of vehicles with some of these technologies available for 2016. Most manufacturers however, have disclaimers on their webpages that the driver is responsible for safely operating the vehicle and avoiding collisions and that the semi-automated technology may not work in all situations. The manufacturer is making it clear that liability still rests with the driver.

Volvo has gone so far as to have pedestrian airbags that inflate over the windshield to soften the impact and lessen injuries if a pedestrian is struck by the vehicle. There is also technology to warn of and avoid striking cyclists, and brake in a turn if a collision is imminent. Tesla has an autopilot system that will park itself, drive autonomously within a lane, and change lanes on command. Technology is far enough along that the Highway Loss Data Institute (HLDI) has conducted early tests with crash avoidance technologies that show that forward collision systems, especially those that brake autonomously, and adaptive headlights, which shift direction as the driver steers, show the biggest reductions in crashes.

 Full autonomous technology is more difficult; the vehicle must know all road rules and be able to identify road signs and hazards from stopped vehicles to animals in the roadway. A large concern is if the vehicle able to handle inclement weather. It is one thing for the vehicle to navigate on its own in fair weather, but what about in rain, fog, or especially snow, when the road surface markings are covered? This is an important consideration. Ford has recently announced that by using Light Detection and Ranging (LIDAR) its vehicle can navigate in the dark without lights. With the LIDAR technology, it is not relying on actually detecting paint lines. Generally the vehicles use not only LIDAR but radar and cameras as well, but this test showed that a vehicle can rely solely on LIDAR and navigate correctly.

 Also what about no win scenarios? If a child runs into the street and the only option is killing the driver or the child, which does the vehicle pick? Various manufacturers are looking at the ethics involved in the no-win situation and are consulting with a variety of experts, from engineers to anthropologists. There may also be cultural differences in what the right answer would be in certain situations.

 Volvo is actively trying to make sure an autonomous vehicle never gets into that impossible choice situation, and is going so far as to accept full civil liability for accidents due to its self-driving technology.

 Communication between vehicles is important; vehicle to vehicle technology allows vehicles to communicate with each other in high and full automation systems. Vehicle to infrastructure technology lets the vehicle communicate with the infrastructure itself as to roadways, bridges, turns, etc.

 The technology allows reducing distance between vehicles; breaking can occur simultaneously between vehicles with no extra space needed for reaction time.

 Vehicles can be linked together in trains at the same speed which improves fuel efficiency. However, until large numbers of autonomous vehicles are on the road, such benefits will be minimal. Vehicle to infrastructure connections are similar to vehicle to vehicle connections, allowing for sensing the environment and other vehicles.

 Cyber Issues

 One of the larger concerns is privacy and cyber security. Some testing regulations require a manufacturer to record data thirty seconds before a collision. In order to record data thirty seconds before a collision, the vehicle must be recording data all the time. Could this data be sold to vendors so that the vehicle, knowing that a person visits Taco Bell every Thursday, would be sure to navigate past Taco Bells on a regular basis? Could this be a mini kidnapping? It could end up like certain phone programs, where you opt out of ads for a fee. For a fee the car takes you exactly where you want to avoid temptation, but if you're willing to allow the ads, it will take you past the bakery, jewelry, and grocery store hoping you'll decide to stop in. Could a private investigator use the data to prove to an individual that the spouse was indeed at a lover's house when the spouse had said he was at work? Could a terrorist hack into the systems and have all AVs speed up to ninety miles per hour and turn right? Experts say that is unlikely, because it is one thing to hack into a system and gather credit card or identity information, and another to input instructions for numerous vehicles to all do the same thing at once. The ability to hack into a vehicle and take control is not limited to autonomous vehicles; already a 2014 Jeep Cherokee was hacked as an experiment and the steering, transmission and brakes were able to be controlled remotely by the hackers. Any vehicle with Bluetooth and other wireless connections can be hacked. The hacking of an individual vehicle to cause group chaos would not be practical since one vehicle causing an accident is not likely to cause more damage than the average texting, eating, or other distracted behavior that is causing accidents now. However, if vehicles are connected a single point of failure or someone hacking into one feature could cause a significant number of vehicles to crash at once, wreaking havoc throughout one or many cities. With a connected autonomous vehicle a thief could hack an expensive car and have it show up at his door for his own use. Or hackers could take over multiple vehicles; but this is a global problem, and not limited to autonomous vehicles. Already the ability to unlock doors and start the engine by hacking into the vehicle has been an issue. BMW fixed a flaw in 2015 that would have allowed 2.2 million vehicles to have doors unlocked by hackers. Chrysler also recalled 1.4 million in 2015 for a similar potential problem.

 Shared Vehicles

 Autonomous technology is not only applicable to individuals and their vehicles. Many in the industry believe that vehicle sharing, where an individual can get into a vehicle he does not own and have the vehicle take him somewhere and drop him off, similar to a taxi or rented vehicle, is where this technology will make the most difference. This ability is widely anticipated by the blind and disabled as a much better way to get around than public transportation or relying on friends or neighbors for rides. The ability to call for a vehicle to take someone to the grocery store at any given time, or to an event or appointment without adhering to a set schedule or imposing on others is an enormous benefit to those with any mobility issues. Uber is already looking into the purchase of autonomous vehicles and has plans to open an autonomous vehicle research facility with Carnegie Mellon University. This would eliminate the individual drivers now currently using their own vehicles to drive for Uber, and would allow Uber to start such a service. General Motors has invested $500 million in Lyft, with the goal of creating an on-demand network of autonomous vehicles. Again, with a passenger who cannot drive, the vehicle must be 100 percent accurate in its navigation and driving.

 Private vehicle owners could do the same thing, similar to RelayRides where an individual lists his vehicle on the website with parameters for rental such as cost per day, week, or month and where the vehicle is located. Renters then make a request to use the vehicle; the owner approves it, the renters' driving records are screened by RelayRides. The vehicle is then reserved with a credit card and the owner and renter meet and the keys are given to the renter. With autonomous vehicles, as long as the vehicle stays in autonomous mode the rider's driving record is not much of a concern. If the vehicle is modified so that it did not even have a steering wheel and pedals, then the chances of the rider having an accident are significantly reduced, and the only real concern would be damage done to the inside of the vehicle by the rider. This would allow individuals to use their personal autos to make money on the side by renting them out all day while at work.

 Zipcar is a similar service, although the vehicles are owned by Zipcar and are parked in various locations and the renter picks them up at that location. Autonomous vehicles could be parked at various locations and those who have subscribed to the service could hop into an autonomous vehicle and go. Again, few or no worries about the driving capability of the diver.

 Beverly Hills has announced a resolution to develop autonomous vehicles as public transportation. They would be driverless municipal shuttles available on demand with users requesting rides via smart phone. The plan is to work with AV manufacturers such as Google and Tesla and with regulators and policy makers to draft a white paper outlining the program.

 Liability

 Liability is of course the big question in this issue; just exactly who will be liable if an AV is involved in an accident? Will responsibility fall on the passenger, or will responsibility ultimately fall back on the manufacturer? It is possible that states will develop regulations mandating liability; Florida already has a statute stating that a person is the operator of an autonomous vehicle in autonomous mode if the person activated the technology, even if the person is not in the driver's seat or even in the vehicle. This is important. AVs will have an override available so that a driver can override the autonomous technology; this may disable all the non-mandated safety technology, so the driver is solely responsible for the operation of the vehicle. Liability of course then rests totally with the individual who overrode the autonomous technology.

 Manufacturers may accept liability for collision, especially since courts may hold them responsible for the autonomous technology. They could also rely on product liability coverage for any issues with the programming of the vehicle. It remains to be seen what the state regulators and courts will do as the technology progresses and becomes available. Volvo is already accepting liability for its test vehicles.

 Another issue is liability for vehicles that were not originally manufactured with autonomous technology but were retrofitted after the fact. Google is developing technology to make vehicles autonomous but does not manufacture vehicles; such technology would have to be added to an existing vehicle. However since the autonomous equipment is added to an existing vehicle, if the added equipment fails to work as advertised is Google responsible, is the shop that added the technology responsible, or is the individual in the vehicle at the time of the accident responsible? These and other liability issues may not be sorted out until lawsuits occur.

 Insurance

 Regardless of where liability ultimately lies, insurance is going to be involved. If the manufacturers are held liable for an error by an autonomous car, not necessarily a malfunction, product liability coverage will be invoked. The ISO Products/Completed Operations Liability Coverage Form CG 00 38 04 13 provides coverage for bodily injury or property damage the insured is legally liable for within the defined products-completed operations hazard. The form defines products/completed operations hazard as "bodily injury" or "physical damage" arising out of "your product" unless the work is still in the insured's possession or the work is not completed. Transportation of property causing injury or damage is not covered. "Your product" is defined as goods or products manufactured, sold, handled, distributed, or disposed of by the insured, others trading under the insured's name, or a person or organization whose business or assets the insured has acquired. Therefore, if an autonomous vehicle crashes and the passenger is injured, the product liability coverage could be invoked because the AV technology did not avoid an accident and therefore caused an injury.

 However, what if states or courts decide that the owner/passenger of the vehicle retains liability? The ISO Personal Auto Form PP 00 01 01 05 provides liability to the insured or any "family member" for the ownership, maintenance or use of any auto. However, the policy excludes coverage if the vehicle is being used as a public or livery conveyance. An insured who has signed up for RelayRides and hires out his vehicle would not be covered under the personal auto policy.

 Such services could become models for how autonomous vehicle services could be insured. An autonomous vehicle service may allow passengers to select coverage levels depending upon the amount of risk they are willing to assume. The rider will need individual coverage in case he has an accident that exceeds the limits of insurance provided by the autonomous vehicle company.

 As far as coverage for the vehicles in an autonomous fleet that operates similar to a taxi operation, coverage would be provided by a business auto form. The ISO CA 00 01 10 13 provides both physical damage and liability for the owner of the vehicles.

 The potential impact on the auto insurance industry is huge. Some experts predict that the industry will shrink by sixty percent within the next twenty-five years. With autonomous vehicles as the main vehicle type on the road, accidents will plummet, reducing rates substantially. Allstate has admitted that it does not know how it would survive the spread of autonomous vehicles, fearing that it could disrupt the demand for the standard auto products. Many executives agree with the sentiment, but see autonomous vehicles as a problem that will not develop until well into the future. A significant percentage of the fleet of American vehicles would need to be autonomous for it to impact the industry greatly. As not just autonomous vehicles grow, but the improvement in safety equipment overall reduce the occurrence of accidents, the industry is going to have to face many changes in the future. The possible reduction in those even owning cars because of reliance on demand car-sharing services or autonomous transportation is huge.

 The move to autonomous vehicles may begin with long-haul trucking; the ability to have the vehicles moving 24/7, without drivers, no required downtime due to driver fatigue, better fuel economy makes moving goods across the country with an autonomous fleet an engaging idea. However pushback could come from unions and professional drivers. Drivers do not want to lose their jobs, and with a driver present there is someone there to pay attention to the cargo and help prevent theft. With technology as it stands now and the feelings of regulators, at least the first generation of autonomous trucks would likely require a human driver. Further generations though could become completely autonomous.

State Farm seems to be preparing for a shift in the industry by focusing on life management issues, which would involve analyzing customer data on all risks, including home, auto, and personal health. Patterns within the data could be used to offer services and discounts.

There is speculation that coverage will shift, and not disappear. With the liability issues, there may be a greater demand for D&O coverage for those that make, service or supply parts for autonomous vehicles, as well as liability for those providing the connectivity between vehicles and infrastructure, and cyber liability coverages as well.

 Rating

 One of the advantages predicted once autonomous vehicles take hold is a significant reduction in insurance premiums. While this may eventually happen, there are a number of factors that must be considered. First, it will be many years before autonomous vehicles make up most of the vehicles on the road. The average age of vehicles on the road today is ten years old, so it will be many years before the fleet of vehicles turns over. Not everyone is willing to accept the technology; some people are not sure that the vehicle can really handle all situations, and others simply like to interact with the vehicle. There are still those who drive a manual transmission because they enjoy having more control over the vehicle. An insured's driving record, accident history, and location (urban or rural) will still be important factors in creating premiums for many years.

 Individuals who have ridden in autonomous vehicles report that other drivers are fascinated by the vehicles and some drivers will speed up or slow down to look at the vehicle; this presents a distraction to other vehicles. While the AV may not be involved in an accident, it may have a role in other vehicles having accidents. Likewise, there may be people who drive autonomous vehicles to see how the vehicle reacts; they may want to see if the vehicle really will swerve to avoid another vehicle in its lane or will brake in time to avoid rear-ending a vehicle in front. Autonomous vehicles, while safe themselves, may not be able to avoid the distracted or erratic behavior of all other drivers on the road. While the AV may be able to swerve to avoid hitting another car, the vehicle may hit a tree instead, still causing damage to the vehicle.

 There will also still be hail, tornadoes, floods, and other natural events that will damage a vehicle that have nothing to do with the action of the vehicle in motion. These things will still occur, and while they may not occur frequently, the cost to repair the autonomous vehicles may be higher than current repair costs.

 Regulations

 How states regulate AVs is another major factor in putting them on the road. Currently California, the District of Columbia, Florida, and Nevada have regulations for the use and testing of AVs on their roads. However these regulations are specific to testing and do not apply to the general population since the vehicles are not on the market, although California is working on regulations for private use of autonomous vehicles. While the regulations are similar in many ways there are some variations. See the Autonomous Vehicle Legislation chart for details. Florida has recently passed legislation allowing drivers to buy and use autonomous vehicles on Florida roads. The technology is believed to become available in July 2016.

 The regulations contain various definitions, including artificial intelligence, autonomous vehicle, driver, and manufacturer. Most testing regulations require one or two people in the vehicle, with one person in the driver's seat at all times ready and able to take control if the vehicle malfunctions.

 Insurance requirements range from minimum state financial responsibility requirements to $5 million. Insurance policies, surety bonds, and self-insurance are acceptable ways to insure the driver against liability.

 Requirements for the vehicles themselves include the ability to operate on the states roads, mechanisms to disengage the autonomous technology, ten thousand miles of previous testing, and compliance with all federal and state safety guidelines.

 Some states require reporting of any accidents and some require storing of sensor date for at least thirty seconds before a collision occurs.

 Some states have already stated that an endorsement will be used for test drivers that states that a person is not actively required to drive the AV. Other states simply require the driver to have a valid license. However licensing will be a large issue. Will individuals need training to understand when the vehicle needs a human to take over controls and how to transition from automatic to manual mode? Predictions exists that AVs would be helpful to the elderly, disabled, and others who cannot drive, and that this is one of the major benefits of AV technology. If states require a driver to have a valid license, is the benefit to these non-driving individuals taken away?

 The California Department of Motor Vehicles has draft regulations that would require a licensed driver to be in any autonomous vehicle. Those attending a public hearing on the proposed regulations raised several factors; if the purpose of autonomous vehicles is to not be in control of the vehicle, is it realistic to expect someone to be paying attention to take over suddenly? Disabled individuals were very vocal with their concerns that requiring a licensed driver makes it impossible to use this technology that could be tremendously helpful to them. The Alliance of Automobile Manufacturers agreed and indicated that requiring a licensed driver to be present negates the purpose of being able to have a vehicle drive itself; it stifles the mobility benefits that it was designed to provide. Also, liability would rest with the driver in the vehicle, even if the vehicle is in autonomous mode, and not the manufacturer.

 Another potential issue is the acquisition of driver's licenses; if autonomous vehicles become more prevalent, will it become more difficult to obtain a license since human error causes the most accidents? Will driving become prohibited?

 The U.S. Department of Transportation will allocate $4 billion over a decade into research and development of autonomous vehicles. The plan will look at technology as well as develop a uniform set of safety rules for autonomous vehicles. States will be able to modify the rules, but this gives everyone a starting point.

 On the Street

As technology has progressed more and more companies and states are trying to test autonomous vehicles on the streets. Uber is working on mapping out Singapore streets so that once implemented Uber vehicles can take the shortest routes to various pickup and drop-off points.

 A driverless airport shuttle is being tested in New Zealand at Christchurch airport. At first the tests are on private airport roads and then move onto public roads after a two year trial period. Because the area has been prone to earthquakes, the vehicles have had accelerometers installed in order to detect abnormal or unusual movement or debris and stop immediately.

 GM has been testing autonomous vehicles on streets in California, Arizona and Michigan and is encouraging other states to do the same. While the infrastructure needs work, GM is not waiting for infrastructure changes and it working to make the technology work with existing roadways. Tesla states that all of its new vehicles will have the hardware necessary to be fully autonomous by October 2017.

 Germany and France plan to test autonomous vehicles over a seventy mile stretch of road between the two countries. The route will allow testing of 5G wireless communications between cars and infrastructure, vehicle operations such as overtaking, braking, emergency warning and call systems as well as others.

Conclusion

 The possibility of autonomous vehicles is exciting new technology, and the potential benefits to society and individuals is enormous. However there are many issues that need to be resolved, many of which cannot be solved in advance of states creating legislation or even the first lawsuit to determine liability when an AV has an accident. Infrastructure also needs to be addressed, because technology for the vehicles to be able to understand a roadway covered in snow may ultimately be required. Changes to infrastructure are expensive and time consuming. However, beneficial changes to infrastructure are possible as well. Optimized use of city parking and reduced lane widths which could improve traffic flow and reduce congestion are just a start.

 Other areas will be impacted as well; with many organ donations coming from fatal auto accidents, where will organs of the future come from? There is already a shortage, will the saving of lives on the highway result in deaths of others from lack of available organs? With reduced injuries from accidents, how will hospitals be affected? There are plenty of illnesses and accidents to be sure, but will there be a shift of some sort in emergency medicine?

 Pollution could be reduced as autonomous vehicles improve fuel economy which could benefit the environment and society as a whole. With traffic density increased the required road space would be less, also positively impacting the environment.

 There are many, many issues involved in the development and use of autonomous vehicles. This is a rapidly growing and changing technology, and while it has its naysayers, many think this is the future of automobile traffic.