This chapter examines some of the main factors affecting public acceptance of driverless vehicles. It first discusses the existing research on this question, with surveys and studies on attitudes towards driverless vehicles in Australia. The chapter then outlines the major benefits which driverless vehicles are expected to bring before focusing on some of the factors which have been identified as the key barriers to public acceptance. The chapter concludes by emphasising the importance of public engagement in building acceptance of driverless vehicles.
Surveys and studies on attitudes
Given the relatively new nature of driverless vehicle technology, only limited research has been conducted to date on the attitudes of Australians towards these technologies. The Department of Infrastructure and Regional Development (DIRD) pointed to a 2014 survey conducted by researchers from the University of Michigan of 500 Australians, which found that:
A majority (61.0 per cent) of Australians were aware of ‘self-driving’ vehicles, with a similar number (61.9 per cent) having a positive general opinion.
Sixty-seven per cent of Australian respondents expressed an interest in having automated vehicle technology and that 25 per cent stated they were willing to pay more than $3 000 for highly automated capability. Around 30 per cent of respondents would be unwilling to pay anything extra for an automated vehicle.
A majority of respondents expected better safety, cost of insurance, fuel consumption and environmental outcomes (but not shorter travel times or reduced congestion).
A majority were ‘moderately’ or ‘very’ concerned about some aspects of automated driving technology, including:
system failures (including safety and security);
riding a vehicle with no driver controls;
automation of commercial vehicles and public transport;
automated vehicles getting ‘confused’;
unoccupied trips by automated vehicles;
interactions between automated vehicles and vulnerable road users; and
A significant percentage would watch the road even when not required (43.4 per cent) or would not ride in an automated vehicle (21.2 per cent).
In 2016, the Australian Driverless Vehicle Initiative (ADVI) conducted the first national survey on Australians’ attitudes to driverless vehicles. That survey found that, while many Australians have heard of driverless vehicle technology, less than 10% reported having driven a car with any capacity to drive itself in some situations. For most Australians, highly automated vehicles are an unfamiliar technology.
Professor Simone Pettigrew provided to the Committee an overview of as-yet unpublished research findings from a survey of approximately 1 500 Australians on their attitudes towards autonomous vehicles:
Overall, favourability was quite high. It was 3.2 overall on a five-point scale. So it was definitely towards the positive end. Males were slightly more positive than females; younger people were slightly more positive than older people. But definitely not any kind of large segregation that we may have expected on demographic factors. About half of the respondents indicated that they would be willing to pay more for a vehicle that was autonomous. They are probably underestimating the cost, though. The average willingness to pay was about $5,000 additional cost to a vehicle.
It should be noted, however, that the majority of existing literature primarily deals with the opinions of Australians towards driverless vehicles in the absence of any direct, firsthand experience of them. As the submission from the Queensland University of Technology (QUT) argued:
There are a growing number of peer-reviewed studies where the member of the public is asked about their acceptance and willingness to use automated cars. Although the questionnaires used in these studies are based on well-established psychology theories, these studies remain highly subjective. The major drawback of such studies is that participants viewed and responded to vignettes or scenarios involving automated vehicles, or descriptions of automated vehicles. Thus, these participants have never experience a real driving an automated car (level 3-4). It could be speculated that individuals still perceive driverless cars as “science fiction”. They cannot form an accurate (non-biased) assessment of the true potential of such a disruptive technology without actually driving it.
As multiple witnesses noted, while only limited numbers of Australians have had the opportunity to experience driverless vehicle technology, Australians tend to be early adopters of many new technologies. This suggests that, as availability of, and familiarity with, automated vehicles grows, and as potential concerns are addressed, Australians may embrace the benefits of highly automated vehicles.
Expected benefits of driverless vehicles
Advocates of driverless vehicles identified a range of potential benefits. This section discusses the major benefits expected as a consequence of increasingly automated vehicle technology. Two of the major benefits identified, improved safety and increased access and mobility, are discussed separately in this report, in chapters three and six respectively.
Use of time
A key benefit identified is that, by removing the driving task, passengers in driverless vehicles will be able to spend their time more productively, enjoyably or usefully. ADVI’s survey found that more than half (56 per cent) of respondents identified this as a potential benefit of driverless vehicles. The survey conducted by the Royal Automobile Club of Western Australia (RAC WA) had a similar finding, with ‘more productive and efficient use of travel time’ as the third most identified benefit.
The submission from the iMOVE CRC noted this potential benefit, highlighting that it could have positive impacts on health, employment and social lives:
Removing the need to focus on the driving task frees up the time in the vehicle to be used more productively for work, relaxation, or even sleep. This would reduce the time cost of travel, reduce the pressure to live close to the location of the work, and more generally support decentralisation. It would also increase people’s geographic range of work opportunities.
Not requiring a licensed driver might also offer families more flexibility as they try to juggle competing demands for getting each member from A to B.
The Swinburne Univeristy of Technology also commented on this factor, and noted that autonomous vehicles will likely result in lowered instances of driver stress.
Congestion and use of urban space
The capacity of autonomous vehicles to decrease traffic congestion has been cited as one of their principal benefits.
Vehicle manufacturer Volvo outlined how this improvement might occur:
In a fully self-driving environment autonomous cars will communicate with each other and the road network via the cloud. This will result in traffic flowing more smoothly, easing congestion on major roads and making these journeys more enjoyable and productive for the driver/occupants. Self-driving cars will be able to merge into traffic and plan ahead more efficiently than those with human drivers.
Connected technology and better all-round awareness means that autonomous cars will reduce congestion on Australian roads, saving millions of wasted hours on the road. Autonomous cars will allow drivers to use their time in the car as they choose – relaxing or working as desired. The car could become an extension of the office and allow commuters to arrive at work less stressed and better prepared.
Professor Hugh Bradlow of the Australian Academy of Technology and Engineering also pointed to this advantage of driverless vehicles, arguing that the fully automated land vehicle system ‘will yield massive convenience factors—for example, congestion will be a thing of the past. Every journey will have near-certain timing, so you will not have to allow for buffers’.
Professor Bradlow also noted that road usage will change significantly with driverless vehicles:
By the way, the autonomous vehicles will not leave four car lengths [between vehicles], because they have reactions of sometimes milliseconds as opposed to seconds like human beings. So one of the savings you get—and we have modelled this on the freeways—is that you can just pack the cars in really tightly, like a train, so you double the lane capacity of the roads.
Improved urban planning and use of space has also been highlighted as a key expected benefit of autonomous vehicles. For instance, as Telstra’s submission outlined:
AVs do not need permanent parking spaces, are likely to be shared so there won’t be as many, and can move away from the urban core when needed. This will drastically reduce the need for carparks and driveways, giving us scope to redevelop these areas for greater social amenity and economic benefit.
The University of the Sunshine Coast submission made a similar point, noting that the likely reduction of designated parking spaces in urban environments could lead to better use of that land. Mr Steven Harrison of the Council of Capital City Lord Mayors noted that this change will be seen in larger cities and is beginning to happen already:
They are already beginning to see in the US changes in the use of homes and apartment construction where there is no car parking required or people are converting their garages into bedrooms or Airbnbs. So those sorts of changes are already happening when people are thinking about design—'Do we really need to incorporate as much car parking as previously?'
We have a couple of apartment projects in Adelaide, and they are very common in Melbourne, where the apartments go up but there might be three or four cars that are shared by the owners of the apartments. So the building actually owns the motor vehicles. You could see there would be a natural progression into autonomous vehicles. They would be autonomous and you would not own them but the building might own them. So we are starting to see those changes. We are experiencing that in Australia.
DIRD also pointed to the potential benefit of improved liveability in cities, noting that it is dependent both on the dominant usage model of autonomous vehicles and active planning decisions:
If automated vehicles do not generate increased travel, then potential improvements in congestion, sustainability and accessibility will serve to improve the liveability of Australian cities and communities. Some analysts envision a future where city structures are transformed and public spaces become cleaner and safer, leading to increased social connectedness and enhanced societal wellbeing.
Reduced requirements for parking space could create surplus land for higher value urban redevelopment and community use. The realisation of these social and structural benefits will require proactive management through changes in land use planning policy and regulations.
Some witnesses suggested that improved transport options will instead see increased urban sprawl as people take advantage of the convenience of autonomous transport to live further away from their workplaces than they currently do.
Alongside the anticipated benefit of easing congestion is the likelihood that driverless vehicles will have a positive effect on the environment. While the main impact is likely to come from electric, rather than autonomous vehicles, the Committee nonetheless heard that autonomous vehicles should result in improved environmental performance.
As DIRD’s submission noted, Australia’s transport sector is responsible for 16 per cent of the country’s total greenhouse gas emissions, with light vehicles making up 10 per cent of the total.
Telstra pointed to the advantages through improved, and more consistent, driving as well as using information available to vehicles via network connections to improve fuel usage.
Volvo made a similar point, noting that autonomous vehicles are likely to reduce the forms of driving which are most fuel intensive:
Autonomous, connected cars will be able to drive more efficiently, reducing fuel consumption and harmful emissions. Better anticipation and communication with other cars will reduce stop/start traffic and heavy braking, and they will be able to form safe, tightly packed ‘road trains’ that reduce aerodynamic drag at speed.
The latter point links to one of the anticipated benefits of truck platooning, likely to be one of the earliest functions of automated vehicles, as Telstra’s submission explained:
AVs in the Transport & Logistics sector can take advantage of techniques like truck platooning (where a truck travels close enough behind another to benefit from the wind break created by the leader, informed by C-ITS) or vehicle to infrastructure (V2I) communication which can provide green light priority for heavy vehicles to reduce carbon emissions from both fuel burn and reduced congestion… For example, Platooning vendor Peloton estimates a 7.5 per cent fuel reduction from platooning just two trucks.
Volvo also noted that self-driving vehicles could have the further impact of making electric vehicle use more efficient, by automatically driving to and from a charging station as convenient, rather than requiring a human driver to do so.
However, as with anticipated reductions in traffic congestion, the dominant use model for driverless vehicles will be a critical factor. Environmental benefits may be lessened if autonomous vehicles lead to a rise in vehicle usage because of the convenience they offer or because they frequently drive without passengers.
Key barriers to public acceptance
This section discusses some of the key barriers to the public acceptance of driverless vehicles as identified by submitters and witnesses. Two of the major barriers, questions of legal liability and the impact driverless vehicles will have on employment, are discussed separately, in chapters four and five respectively.
Data security and privacy
A barrier to public acceptance of autonomous vehicles is the level of concern regarding data issues, particularly in terms of the privacy of personal information and the vulnerability of data to cybersecurity threats.
Privacy, data use and ownership
The submission from the iMove CRC team outlined the concerns regarding privacy and vulnerability to hacking:
People are justifiably concerned about the collection of data and its subsequent use. We need to have a clear understanding of what data that is being collected in order to predetermine how it is used and ensure that people’s fears are allayed. A global data governance model that considers all information sources would play an important role here for the ‘global good’, however there are considerable challenges in setting this up.
The scale of the data which will be created by autonomous vehicles was explained by Professor Butler from the Queensland University of Technology:
These vehicles are going to generate an enormous amount of data. I saw a quote from the CEO of Intel that suggested that, over an eight-hour period of driving, one of these vehicles will generate and consume 40 terabytes of data. So for every eight hours they will create 40 terabytes of data. That data is very wide ranging, and the implications of it from a privacy perspective may depend on the particular model that we are talking about...
There are issues there about who owns that data. The manufacturers may lay claim to that, and likely will lay claim to that sort of data. To that end, I saw an announcement by Ford that they are spending $200 million to convert an assembly factory in Michigan to a data-processing facility. So they are certainly planning on collecting a large amount of data.
The Federal Chamber of Automotive Industries suggested that autonomous vehicle data could be broadly categorised into three types and noted that each of those will need to be considered in different ways:
Vehicle owner/driver information (including location data); and
Vehicle systems operation data.
As Professor Butler indicated, there will likely be consequential use of that data too, as the companies who own it find ways to commercialise the information they collect about vehicle operators:
The vehicles will record locations that you might like to frequently visit. If you like to go to a particular fast food restaurant… you might find then that, if the manufacturer is collecting that sort of data and passing it on to other fast food chains, there are implications of being deluged with advertising from rival food outlets.
Furthermore, there are even more concerning aspects about access to that data:
So these things are recording where you have been. There is a question there about who might access that information, such as whether, if you have a suspicious partner who is wondering where their partner is going to, that person can access the information. That might raise concerns about domestic violence, if that person is of a certain inclination.
The Department of Industry, Innovation and Science (DIIS) described some of the effects of cybersecurity lapses as seen in the United States:
In July 2015, Fiat Chrysler recalled 1.4 million vehicles in the US to address a software vulnerability in its systems, following a reported demonstration that allowed security researchers to remotely hack into and control a Jeep Cherokee’s engine via its internet-connected entertainment system. The New York Times also reported that the same researchers demonstrated a way to control hundreds of thousands of vehicles remotely from the internet: they were able to track cars by their location, see how fast they were going, and control lights, windshield wipers, navigation and in some cases, brakes and steering.
The iMOVE CRC submission also discussed the problem:
… there is considerable concern about the perceived vulnerability of individual vehicles and the traffic system as a whole. There is potential for many undesirable scenarios if the systems of driverless vehicles are compromised. These concerns are magnified by the high level of connectivity between vehicles, and their integration with myriad other devices that could introduce malware or spyware.
However, Mr David Pickett of Volvo Australia clarified that a key cybersecurity concern – that of a system being hacked to gain control of the vehicles reliant on it – will not be a problem, since there will be no capacity to control vehicles in this way:
There is not that much communication allowed back into the car. For someone to get in and start controlling something is quite a different design to what the car is set up to do. It is not designed to take that sort of input. For majority of the steering and braking control, the car is controlling itself and what it sees. The only information it takes is that, maybe, the Sydney Harbour Bridge is closed because of an accident, so it would divert using the navigation system… The systems are not designed to allow that technology to access the car. You do not have access into the vehicle from outside.
The Government’s response to such concerns includes the $230 million National Cyber Security Strategy, which includes the Cyber Security Growth Centre:
The part that [the Department of Industry, Innovation and Science] is responsible for is the Cyber Security Growth Centre, and they have developed a sector industry plan. Cyber security obviously covers lots of different issues around national security, but the growth centre is looking at the business opportunities and industry demand. So they are looking at technologies around privacy, trust and security, particularly in relation to things like cloud computing but also autonomous vehicles, robotics, the Internet of Things. This is to make the point that we have been making to the group during our discussions, which is that the government is working towards building the framework for all of the aspects—whether it is a satellite positioning, the regulatory structure, the cyber security—for what is going to be a 10-15 year implementation of autonomous vehicles. It involves looking at the skills, doing a lot of the structural stuff for the economy.
As DIRD noted, Australia’s place in the vehicle market means that its policies align with international standards and markets. As such:
The Department is engaging with international bodies who are developing standards and guidance for automated vehicle cybersecurity, such as the World Forum for the Harmonisation of Vehicle Standards. Separately, collaborative work is underway with state and territory governments to develop a security management plan for connected and automated vehicles, focusing on the security of wireless communications between vehicles, and with roadside infrastructure.
Enjoyment and cultural issues
As multiple witnesses noted during the Committee’s hearings, for many Australians vehicle ownership and driving are not exclusively functional. Many people enjoy driving their car and consider it an important cultural marker. For that reason, the prospect of having only limited – or even zero – control over their vehicle is a prospect to be dreaded rather than positively anticipated.
For instance, Dr Hsu of the ARC Robotic Futures Team noted that ‘a small but not insignificant body of research which finds that there is a cultural tendency in Australia today to equate private car ownership with personal autonomy.
iMOVE CRC noted this point, highlighting that Australians’ ‘love affair’ with their own vehicle is a strong cultural thread’, which will ‘need to be acknowledged in any plans for the introduction of new technologies’.
The Australian Academy of Technology and Engineering’s Professor Bradlow suggested that a useful way of combining the benefits of a fully automated vehicle system with people’s desire to drive themselves could be found via analogy with the last great shift in land transport:
The world made a very successful transition at the beginning of the 20th century from horses to cars. It took about 10 years, in fact. People still like to ride horses; they just do not do it on the Monash Freeway. The same thing will apply to cars. They can go out to Calder Park and race their cars or drive their cars there.
As has been noted, a primary factor affecting the level of public acceptance of driverless vehicles is the unfamiliarity and lack of experience most people have with them. As a consequence, the general public’s understanding of the currently available level of technology, and realistic expectations of how and when this will further develop, is often inaccurate.
One analogy the Committee heard compared the development of autonomous vehicles to that of lifts:
… when lifts were invented: you needed an operator; then there was a point where you still had an operator but it was actually automated because people did not accept it or trust it; and eventually there were no lift drivers and now they are fully automated. It could have happened a lot earlier except for acceptance. It is really important that the government leads on outreach and social acceptance of driverless because with that acceptance and the real world experience of seeing these trials—and seeing is believing, if you like—then people will want to be involved in it. That is when you get that acceleration of benefit.
Telstra made a similar argument in its submission, noting that, ‘to achieve the required level of social acceptance, reliability needs to be demonstrated through pilots and public participation’.
As Dr Wenham of the Australian Academy of Technology and Engineering argued, it is vital for there to be public engagement on the social issues around driverless vehicles, noting that with new technologies, social licence is key:
Pick your technology; the issue is rarely with the technological aspects—that will be resolved with research that is going on in markets and that sort of work—it is around social acceptance and social licence. If the conversation is not structured properly with the community and people do not understand the issues around this and are not able to feel that they can have a say in how these technologies are deployed, you will have the sort of problems that you had with other technologies. We should not underestimate that social licence.
Inevitably, as both the technology itself and its spread improve, the attitudes of Australians towards driverless vehicles will change. The Committee notes that the National Transport Commission released its Guidelines for Trials of Automated Vehicles in Australia in May 2017, and expects that as trials of vehicles with autonomous capabilities expand, some of these concerns will be alleviated and benefits will be recognised. However, as with all emerging technologies, the Committee notes that increased use may also identify further problems to be resolved. Similarly, additional benefits may also emerge as the adoption of highly automated vehicles become available.
Having heard about some of the primary concerns and barriers to public acceptance of driverless vehicles, the Committee emphasises that all stakeholders involving in introducing these vehicles need to engage with community concerns and ensure that these are taken seriously. Driverless vehicles have the capacity to bring about substantial social benefits – including those discussed in this report – but without adequate public engagement, those may never be fully realised in Australia.
As part of that public engagement, the Committee is of the view that the Commonwealth Government should continue to ensure that addressing concerns in regards to data privacy and cybersecurity forms a part of Government’s preparations of driverless vehicles.
Overall, the Committee found that the main issue affecting public acceptance of driverless vehicles is the lack of familiarity most Australians have with vehicles containing high levels of automation. Given the substantial safety benefits, improved access and mobility options, better use of time and improved environmental and planning outcomes likely to result from the availability of driverless vehicles, the Committee considers it to be of primary importance that there is genuine engagement with Australians’ concerns about and inexperience with driverless vehicles.
For that reason, the Committee considers that the Commonwealth Government should facilitate and encourage public trials of driverless vehicles to increase public understanding and develop further dialogue on any future social issues.
The Committee recommends that, noting the range of benefits automated vehicles are likely to bring and the need for public acceptance of the technology, the Commonwealth Government facilitate and encourage trials of automated vehicles in Australia, with a particular focus on trials that enable members of the public to experience automated vehicles on public roads.
The Committee recommends that the National Cyber Security Strategy specifically investigate automated vehicles (and associated transport systems) to address potential vulnerabilities relating to automation.