Autonomous cars

STC 100 David Mercer

Cars/ Driverless Cars and the politics of innovation

Visions and Imaginations

• Like Digital and Personal Medicine the possibilities of driverless cars have captured the imaginations of governments and entrepreneurs so we need to be aware of hype re: the possibilities and Utopian and Dystopian Imaginaries of autonomous cars

• We also need to remember they form part of broader socio-technical systems and that the traditional car and transport systems they have been linked to have long histories and that different actors may have different visions for the future of AV’s. These factors will exert an influence on the future direction of AV’s.

• So before discussing AV’s some general history of cars.

History: Inertia and Politics and Socio- technical systems

• The broad history of innovations in cars/automobiles has featured incremental improvements in efficiency and safety but a lack of revolutionary changes in design

• Patterns of resistance to change, not only on the part of designers but users and regulators.

• Arguably one of the most conservative technologies of the last 100 years! Comparable with Coal Power!

Case studies of resistance to innovation: car safety and electric cars

• Cars have been around from the early part of the 20 century (and even earlier if we include motorised carriages)

• They became one of the first important mass produced products linked to developments in the early 2oth century factory system. Eg Fordism

• Mass markets for auto’s particularly boomed in the post World War II period with large increases in ownership by the average family. Cars became faster and highway development expanded across much of the developed world following a largely US model

• By the 1960’s- growing awareness of large number of serious accidents and fatalities

• Some concerns to increase the quality of driving skills and also reduce things like drink driving

Designing safer cars?

• Manufacturers and Engineers also began to consider problems in the ways cars were designed even low speed collisions often resulted in fatalities

• The idea of seat belts already used by pilots and redesigning the interiors of cars to reduce sharp objects and more carefully locate fuel tanks to avoid explosions were all explored and tested....and demonstrated to be effective by engineers.

Resistance from Manufacturers

• Car Manufacturer’s systematically (with few exceptions) resisted these initiatives to introduce more safety features into auto’s

• They argued that safety devices went against the free- spirit of car ownership and driving which went with the time and didn’t suit their marketing strategies

• They also suggested that car safety was first and foremost a question of the quality of the driver not the car (technology)

• Actively suppressed engineers and debate about the suitability of auto design and investment in safety technology

Ralph Nader and the birth of the consumer movement’

• Lawyers for car accident victims and crash investigators in the United States became increasingly troubled not only about the design of cars and safety but the fact that manufacturers appeared to know about dangers but avoided doing anything about it.

• In 1965 Ralph Nader a lawyers published a book titled ‘Unsafe at any Speed’ . Nader documented the lack of corporate interest in fixing known design failures such as Chevrolet Corvair 1960-63 which had faulty suspension which lead to cars flipping over and numerous other models where drivers in minor accidents were being impaled by steering wheels and sharp car interior trimmings

• When Nader started publicly pursuing these claims he was subject to harassment and stalked by thugs paid by the auto-manufacturers

• Ultimately Nader was successful in having the matter dealt with by a US congressional Hearing and the automobile manufacturers were compelled by government to start acting on car safety.

Death vs Dollars • Nader further pursued his challenge to auto companies with legal

actions during the early 1970’s involving the Ford ‘Pinto’. • The Ford ‘Pinto’ was extremely vulnerable to bursting into flames

with even minor rear end collisions because of a design flaw. • Leaked documents showed that Ford’s engineer’s were aware of

the problem but that the company calculated the cost of recalling the vehicle was more of a financial risk and possible burden than dealing with the cost of accident claims. Eg: Death vs Dollars

• Numerous legal actions which followed cemented into US law the need for auto-manufactures to take building safety into car design more seriously.

• At the time of the publication of Unsafe at any Speed in 1965 there were 5 deaths per 100 million miles travelled in cars in the US in 2014 only 1.

(See C.Jensen, ‘50 Years Ago, Unsafe at Any Speed’ shook the Auto World’ NY Times, Nov 26, 2015)

Resistance from Users

• Even as safety devices were increasingly introduced into car technology users were not always willing to pay extra for optional improvements.

• Expensive car-makes such as Volvo used to use their extra safety features a marketing point but this didn’t become a dominant theme in all auto-marketing where speed, fashion and lifestyle still key themes.

• Mandated use of voluntary safety devices such as seat belts was slow to be introduced in some countries, particularly in some parts of the US where seat belts were seen as an attack on personal liberties.

• Issues of car –responsibility for safety as a user vs technology question persist as a feature of car safety debates and re-appears with AV’s

(for more history see Wetmore reading in subject outline)

Culture of Car Risk

• How people drive also influenced by culture. When you drive a car do you become something more or less than a person as you take on the identity of being a driver.

• A ‘driver’ is part-human-part machine (a cyborg ?)and part of a road system or ‘form of life’ eg: when we see a car on a highway its expected but a person walking down the middle of a large highway is ‘unnatural’

• Drivers adopt risk behaviours consistent with the system they are part of which includes road rules, training and the behaviour of other drivers.

• Drivers may take more risks in cars that are designed to be safer.

Resistance to Innovation 2: Electric Cars

• Documentary by Chris Paine 2006 (available in the library) Who Killed the Electric Car documents how for more than 15 years political and social obstacles starting from the early 1990’s were placed in way of early attempts to promote electric cars in California.

• The documentary demonstrates that these factors were much more significant than any straightforward technical limitations of electric cars

• Electric cars are becoming more common across the world but its taken close to 40 years and they are still far from being the main form of automobile

Why electric cars ?

• Electrically powered cars were explored as a possible alternative to petrol driven cars in the early years of car development but petrol driven cars predominated.

• During the 1970’s and 1980’s increasing concern in major cities with high levels of smog (air pollution) linked to automobile engine exhaust emissions which were in turn linked to increased levels of serious lung diseases.

• One of the world’s worst cities for smog was Los Angeles in California in the US.

• Concerns lead to medical organizations and health authorities to lobby US Federal governments to draft laws to try to reduce vehicle emissions. This was largely ignored at this level of government so the Californian government introduced its own legislation to try to force manufacturers to start reducing vehicle emissions

Zero Emissions Vehicle Law

• California is the world’s 8th biggest economy so this was a regulatory initiative with implication across the US and the rest of the world.

• Concerted attempts to develop the electric car in California were linked to this so called 1990 Zero Emissions Vehicle law.

• There were also Important engineering initiatives at the time particularly at General Motors also encouraged interest in promoting electric cars

GM and the EV1

• During the late 1980’s GM had instructed some of its senior engineers and marketing people to look into the development of electric vehicles

• The project initially took off and the task of producing an electric vehicle for mass production was delegated to a small group of scientists, engineers and designers at its Technical Center in LA

• By most accounts by 1995 they had managed to produce a high quality high performance car for the market

• Between 1996 and 1998 GM produced 1,117 EV1’s 800 were made available on 3 year leases

• Although the company didn’t allow full purchase those who leased them provided very positive feedback

• Designers and Engineers reported that they believed they had come up with a successful breakthrough technology

Stalling the electric car?

• By early 2003 the California Zero Emission Vehicle law was overturned (although it still exists in a modified form) and electric car development stalled.

• Despite enthusiasm on the part of Engineers senior management at GM and marketing didn’t appear to promote the technology in ‘good faith’ marketing campaigns were unusual, EV leaseholders were prohibited from buying their cars and all EV’s on lease were recalled and destroyed in 2005

• GM executives fought the ZEV law and put considerable effort into taking over the company that produced the Hummer which became its flagship vehicle instead of Electric Vehicles

• The rights to superior Battery technology which was first bought up by GM was sold to Exxon-Mobile and EV’s not fitted with the best batteries possible.

A slow journey

• In the meanwhile the Fossil fuel industries had been undermining the electric car initiative by supporting groups who opposed legislation the Californian government was trying to introduce EV recharging stations throughout California

• Rather than promote the EV initiative US Federal government in response to the Fossil Fuel lobby instead put investment into more speculative Fuel Cell technology for cars which was far less well developed than EV’s and therefore far less likely to place any immediate pressure on the profits of fossil fuel industries

• Electric cars and hybrids are now becoming more common with recent growth in numbers suggesting the beginning of an important shift but it has been a slow journey with plenty of politics

MOVING ON: Autonomous Vehicles(AV) AV’s and Interest Groups and Visions

The history of innovation in car technology discussed above: •(1) Reminds us to consider the tendency for inertia and entrenchment in car and related transport socio-technical systems. This is quite unlike the relative rapid developments in the adoption of consumer electronics and information technology over the last 20-30 years. • (2) To make sure we don’t just look at the possibilities that the technology might offer but also the competing visions that different engineers, governments developers and possible future users have about the future of AV’s.

Autonomous Vehicles as inevitable ?

• Organizations promoting autonomous vehicles regularly suggest AV will inevitably be adopted. Drift into soft Technological Determinism. We have little choice but to adapt and accept. Underplays the different forms the technology may actually take in the future, that it might be legitimate to resist or have contested visions for AV’s, or AV’s may have negative consequences for some people

• “ There’s no doubt cars with internal engines will be autonomous but the issue is we need to prepare now if we are going to deal with the challenges automated vehicle present” KPMG spokesman.

• “Our education, advocacy and demonstration efforts help to inform and raise awareness, encourage community acceptance, and promote understanding of the lifestyle benefits of driverless vehicles” ADVI (Australian Driverless Vehicle Initiative)

Key Promoters: Google/Uber and their visions and politics/economics

• “Cashed-up IT disruptors like Google and Uber see new types of vehicles and new patterns of ownership as the basis for new transport economies. They want lightweight, utilitarian “robo-taxis” owned by a corporation and rented by the trip. Travellers will use phone apps or their next-generation successors to do this. This, in the jargon, is “mobility as a service”. These companies’ ambition is to carve out a large niche in competition with private cars, taxis, conventional public transport and even non-motorised transport. Fleets of shared vehicles in constant circulation can reduce the number of individually owned cars and, in particular, the need for parking. In some circumstances, this may support more compact urban forms. But while sustainability or social objectives might be part of the pitch, the profit motive remains dominant.”

( Quoted from: ‘We must plan the driverless city to avoid being hostage to the technology revolution’ Stone, Curtis, Legacy, Sheurer The Conversation. April 28, 2017.)

Google/Uber $

• Hope to link AV’s to areas where they already have a technological advantage expand their influence and economic technological investments into new areas

• ‘Solutions looking for problems’ to attract share-holder investments

• Hook into techno-utopian cultural imaginaries, style of millennials: If you like your iPhone why not your AV ?

• Broader Imaginary of inevitable Automation and De- skilling as economically progressive. Eg GPS devices help de-skill cab drivers assisting the entry of Uber drivers who in turn are de-skilled by AV’s ….

Many Governments have also caught the AV bug

NATIONAL BREAKING NEWS

$10m for more driverless car trials in NSW

Australian Associated Press

9:18AM June 18, 2018

“ NSW could soon be home to more driverless cars, with the state government setting up a special $10 million fund for more trials of the technology.

The kitty, to be announced in Tuesday's state budget, will allow governments, universities, the private sector and start-ups to join forces and test driverless technologies in NSW's cities and regions.

"The future belongs to those who hear it coming, and this investment looks to harness the power of technology to improve lives across the state," Treasurer Dominic Perrottet said in a statement on Monday. While it still might feel like something out of The Jetsons, the driverless vehicles are already being built and used around the world”.

Why might many Australian Governments be

interested in AV technology ? . Possible Reasons: •Fear of falling behind in the development of new technologies more generally •In Australia traditional car industry closing down so fewer entrenched interests in traditional auto’s and possibility of new industries? •Infrastructure crisis in things like public transport in major Australian cities especially Sydney and Melbourne. AV’s might offer cheaper options than expanding bus/rail and AV’s possibly run by private business interests eg: Uber so maybe less direct cost to government in the future •‘Trojan Horse’ to promote driverless forms of traditional public transport. The public may resist the idea of driverless cars but by comparison the greater use of driverless busses, trains and trams seem more mundane and acceptable •Deskilling possibilities used to leverage lower wages for train-tram drivers or replace them.

Not all governments and countries share the same visions

• Different legal systems may encourage or hinder AV development eg:in the US longer traditions of litigation re: accidents and risk than countries with stronger central regulation. For example if the Chinese Govt wanted to introduce AV’s easier for them to control the legal environment .

• Some European and Asian countries/cities already long traditions of driverless trains and trams and densely populated cities with lower levels of individual car ownership than countries like Australia and US.

• Sprawling US and Australian cities with high levels of individual car ownership means replacing existing cars with AV cars much more disruptive

• Some developing economies car ownership low but potentially huge with growing middle classes who may aspire to car ownership Introducing smaller number of shared AV’s might offer an alternative

Surveys indicate important differences in public attitudes to AV’s across the world

(Points from M. Wade Sydney Morning Herald , 6 April 2018 ‘Australia more cautious about driverless cars than many nations: poll ‘ )

•One in six Australians said they would never use one.

•An Ipsos survey covering 28 countries found Australians were less optimistic than the international average about the perceived benefits of driverless cars including safety, speed, efficiency, cost, comfort, environmental impact and enjoyment. (although more positive than the US, UK, Germany and Japan)

•Respondents in emerging economies such as India, China, Malaysia and Mexico tended to be much more enthusiastic about autonomous vehicles.

Age and cultural factors

• 21 per cent of those aged over 50 said they would never use a driverless car compared with 8 per cent of those under 35.

• In Australia 25 per cent said they “can’t wait to use” a driverless car – 5 percentage points lower than average. In China and India nearly half “can't wait” .

• A report by consultancy firm KPMG released in January ranked Australia 14th out of 20 countries for autonomous vehicle “readiness”.

• More Australians would prefer to own an autonomous vehicle (34 per cent) than adopt other proposed models of use including hiring a driverless car on a per-use basis (19 per cent) or leasing one for a monthly subscription fee (9 per cent). China, which has a very different car culture to Australia, had strong support for a subscription model (35 per cent).

• More Australians would prefer Government regulation of safety than Americans who favoured corporate self regulation.

(Wade, Ibid)

What does it actually mean to suggest the emergence of AV’s are inevitable if there are different AV future imaginaries and different interests ? Some Hypotheticals?

A wealthy Uber executive imagines an AV picking them up from their office in San Francisco offering a cheaper option to the Taxi of the past and preferable to mass public transport and something that will generate $ to Uber from renting them to newly formed AV cab companies.

An ambitious Town Planner in Vancouver imagines how AV inspired IT and sensor technology might improve and expand pre- existing driverless trains and free up public space in the city by reducing car transport

An Australian or US Town Planner aware of legal issues surrounding possible risks of AV’s suggests that special lanes are allocated for AV’s only in major cities reducing public space and freedom of movement for non AV owners.

An aspirational middle class bureaucrat in China imagines taking out a shared lease with fellow workers on an AV from the government.

Further possibilities

• An executive from an auto-company imagines how to integrate AV technology into the current model of individual car ownership in the US or Australia. Possibly increasing suburban sprawls by promising driving will take less effort and be safer… so why not live even further from work ?.

• Public Servants advising Australian State Governments imagine AV technology as way to save costs by replacing train and tram drivers

• A person who is old or has a disability which prevents them from having a drivers licence looks forward to having greater future mobility

• Someone who enjoys the experience of driving hopes the technology will be banned.

• How might these different visions and interest shape what happens and how does effect claims that AV’s inevitable. Inevitable in what sense?

AV’s Risk and Safety • High Profile Recent Fatalities involving AV’s have raised numerous

issues about Risk and Legal Responsibility. Are we non-consenting parties to a giant technological experiment? ( see ‘ABC News’ inks in tute readings)

• Various Governments across the world beginning to authorise AV trials (including Australia) questions of insurance/ liabilities and legal responsibilities.

• In Australia currently a driver still needs to be able to control the vehicle.

• Worth noting: The worst health effects of cars not accidents but links to lack of exercise !! Some estimates modelled on Melbourne eg: Disability adjusted life years for every 100,000 people: If 10% of motorised transport was shifted to walking and cycling. Negative -34 because of cycling and pedestrian injuries BUT + 708 for reduction in heart disease and type 2 diabetes. J Sallis The Conversation 4th Jan 2018.

Ethical dilemmas in programing cars in accident scenarios

• German government recently released proposals to design AV’s to operate according to a utilitarian ‘least harm’ principles in the case of probable accidents. Least harm not as simple as it sounds, to put into practice.

• Following these principles the AV would be programmed when faced with the option of killing the group of rowdy drunks veering into the path of the vehicle, to drive, instead into a brick-wall killing the passengers (e.g. mother and their young child) because this would constitute fewer casualties. Could a sensor be designed to differentiate between Drunks, Terrorists and Schoolchildren (probably not)

• Perhaps it could be programmed to discriminate against those encountered on the road who were not obeying traffic laws?

• ‘Finer grained’ ethical questions matching consequences to blameworthiness eg: the negligent or malicious drunken crowd vs the innocent mother and child may be difficult to deal with in programming AV’s in relation to possible accident scenarios.

Limits on machine learning?

• AV’s rely on machine learning (Artificial Intelligence)which gathers huge amounts of data to construct algorithms which tune the cars to respond to what is picked up by their sensing devices. History of AI (expert systems)suggests most successful AI has been when the environments the system operates in can be manipulated to make them as predictable as possible

• The possibilities of Hacking could make AV’s vulnerable to uncertainties and breakdowns in control.

• Some environmental contexts may challenge technical limits of AV eg: designing algorithms to avoid Kangaroo’s very difficult.

• Will sensing devices always operate properly in all weather conditions

• AV’s will operate alongside Non AV-drivers and pedestrians who may still act in new unpredictable ways.

Need to consider social learning that goes alongside machine learning

• Making the world predictable to assist AI (and expert systems) also includes re- configuring users to fit the systems (social learning to go with the machine learning). Think of how many conventions, rules, and discipline surrounds flying people successfully in and out of airports. How much do you change your behaviour to suit the AI system you encounter on your iPhone to pay bills etc.etc.

• Business models of UBER and GOOGLE may be limiting the capacity for improving AV systems because knowledge gained from their trials are seen as valuable IP and protected commercial information. This might limit social learning about AV’s and accidents. Some have suggested that AV’s need to have something like the Black Box of airplanes in them and information about accidents shared. (See Stilgoe reading for tutes)

• Creating conditions for AV’s to safely operate might lead to pressures to change the ways roads are designed, traffic rules are constructed, the way users behave and their perceptions of what are acceptable risks.

• So for AV’s to be widely adopted in some form there will also be a need for ‘social learning’ about AV’s and shifts in the culture and regulation of cars and driving.

Overview: Extracts from Pro’s and Con’s of AV from the US Insurance Industry

(Top 20 Pro’s and Cons associated with Self-Driving Cars, Auto Insurance Center. US Industry News and Information site)

•Since 81 percent of car crashes are the result of human error, computers would take a lot of danger out of the equation. •Computers use complicated algorithms to determine appropriate stopping distance, distance from another vehicle and other data that decreases the chances of car accidents dramatically •There are no opportunities for a computer to be "distracted", which is a leading cause of accidents in the United States at present. •Drunk driving incidents will decrease •Police officer focus could be shifted from writing traffic tickets and handling accidents to managing other, more serious crimes

.

.

.

Pro’s Continued

• The U.S. Department of Transportation assigns a value to each human life: $9.2 million. There would be a significant cost savings in many different venues like insurance costs and healthcare costs associated with accident recovery.

• As an article from Forbes points out, there is also a cost savings associated with time. When a computer takes over the driving responsibilities, drivers can use that time to do other things.

• According to Eno Transportation, self-driving cars in large number participate in platooning, which would significantly improve traffic conditions and congestion. This would help to reduce commute times for drivers in high-traffic areas but also to maximize on gasoline usage.

• Less parking structures and parking headaches would be required, since your car could actually drop you off and locate a parking space farther away

• Larger cities are plagued with the problem of providing adequate public transportation. AV’s a new alternative.

• Disabled individuals, could reap the benefits of self-driving cars with new freedom and enhanced mobility, as suggested by the New York Times.

Cons of AV’s

• Just having the ability to operate a self-driving car would require an education on the driver's part, according to Teletrac. While the computer takes over once the vehicle is operational, the driver would still be required to maintain some knowledge about how to operate it safely.

• The cost of implementing the new technology could be way out of reach for most Americans. Currently, the engineering, power and computer requirements, software, and sensors add up to more than $100,000.

• The most savings in terms of cost, time, and lives is going to come from when more people "opt in" to the service. If self-driving cars are not adopted widely, accidents can and will still happen.

• The very security behind self-driving cars would be a major obstacle, especially because the technology would be of very high interest to hackers, as pointed out by the Guardian.

• In order for a computer to operate a vehicle, a lot of information would have to be stored on the software providing an opportunity for a computer built into the self-driving car to collect personal data.

• Self-driving cars would eliminate many jobs in the transportation sector, especially when it comes to freight transportation and taxi drivers. This could have a negative impact on the unemployment rate and the economy.

• A self-driving car doesn't completely eliminate the likelihood of a car accident. In fact, there's no legal precedent for how a case would be handled. The difficult question of who holds responsibility in a car accident- the driver? The car manufacturer? The software developer?

• The cars are not able to operate at a high level of safety in all weather conditions. In fact, heavy rain can do serious damage to the laser sensor mounted on the car's roof, calling into question what role the driver might have to play in the event the technology fails.

• If other technology fails, such as traffic signals that the cars rely on, there's no accounting for human traffic signals. In the event of an accident, for example, where a police officer is directing traffic, the cars cannot interpret human signals.

• The reliance on technology could mean that over time, drivers are no longer equipped with the skills to operate cars. In the event of a technology glitch or recall, drivers might be helpless to get around, having been "out of practice" in the driving world for some time.

Cons Continued • The success of self-driving cars currently relies on accurate mapping

systems through GPS. As anyone who has been advised to turn down a one-way street or been told by their GPS they were driving on a non- existent street can attest, GPS devices are not always accurate.

• The NHTSA remains skeptical of the technology behind self-driving cars, even calling for a ban on them at one point until further testing could be completed.

• Since Google is currently at the helm of development for the cars, other auto manufacturers might sell fewer cars in the event that Google's version takes off.

• The gasoline industry is likely to suffer because, taking the note of "new and improved" it's likely that the self-driving cars would be electric.

• People who enjoy driving are unlikely to buy into the technology that means they no longer need to focus behind the wheel, so they are likely to keep their own vehicles rather than trading in for a self-driving vehicle.