As the automobile and technology industries have made significant leaps in autonomous vehicle (AV) development, e.g., a level-4 automated driving system was achieved in early 2020, highly automated vehicles (levels 3–5) have attracted attention from both academia and policymakers. Even though academic research on AVs has long been dominated by the fields of engineering and computer science, research efforts regarding the impacts of automated driving systems on human factor aspects, traffic flow characteristics, and fuel efficiency have been conducted for the last two decades [1
]. Recently, the interest in wider social, economic, and environmental implications [1
], especially on a call for policymakers to utilize their roles in governing and regulating AV and its potential risks [5
], is growing as the technology has accelerated. In this way, emerging responses to the risks of AV, addressed by different governments, were compared [9
]. Furthermore, a transformation process that emphasizes inclusivity and democracy has been proposed [10
At the same time, the Klynveld Peat Marwick Goerdeler (KPMG), a multinational professional services network, conducted an International Autonomous Vehicles Readiness Index (AVRI) report for the purposes of opening up dialogue and sharing what different countries have been doing in different sectors as part of their contingency plan. Their report suggested that, from 2018 to 2020, governments around the world have participated in a transformation process, from racing to position themselves as leaders in the filed by enabling the technology to be tested [12
], working on delivering the AVs’ anticipated social benefits while mitigating their potential challenges [13
], to dealing with complex tasks in the implementation process that address risks and opportunities for AVs to operate safely and effectively in our society, as AV technology is entering the period of development maturity [14
Most of the above studies employed the methodologies of literature reviews, policy overviews, and data assessments; at the same time, scenario analyses were also popular in identifying plausible future development and sustainability of AV and formulating policies to cope with the AVs’ potential undesirable subsequent influences, in addition to the positive outlook of smart mobility [15
]. On the other hand, rather than attempting to predict what the future with AV holds, Cohen, Stilgoe, and Cavoli conducted systematic stakeholder workshops to generate insights for AV governance in the UK and formulated questions in four categories: technological and market developments, the use of AVs and reactions to them, the wider impact of AV, and the role of the public sector [6
To better capture how the academic and professional communities have built up our initial understanding regarding emerging issues on AVs’ wider impact, governance, and regulations, as well as risk management of the use of AVs, the related literature was reviewed.
As AV technology has advanced rapidly in recent years, academic and professional communities have begun to pay attention on how it can be effectively implemented into our future from various perspectives: (1) by articulating the need for the public sector’s role on the governance and regulation in response to AVs’ potential risks in addition to discussions on AVs’ wider social, economic, and environmental impact in the earlier wave [1
]; (2) by employing scenario analyses on how different indicators might be deployed in order to achieve a future with a better society [15
]; (3) by comparing and contrasting what different governments are doing in terms of policies and legislation, as well as how different jurisdictions were preparing for the future with a joint effort among different sectors [9
]; and (4) by identifying specific questions related to issues associated with AVs [6
]. This study is among the first attempts to explore the overall picture on how people perceive the importance level and urgency level regarding issues associated with AV, and whether AV is a future trend, from the perspective of a social science researcher, by sorting out issues, developing a questionnaire, and investigating how AI experts and CS/EE majors assessed the importance and urgency levels of these issues in preparation for a future with AVs. Specifically, perceptions toward ten issues (governance, regulation considerations, data privacy and cybersecurity, traffic/environmental/public health benefits, economic and industry, public acceptance, social equity, technology innovation, transportation planning and infrastructure) were collected using 66 questionnaire items. Data were collected from respondents of 31 AI experts and 335 CS/EE graduate and undergraduate students in Taiwan in April 2020. The findings regarding each research question were discussed as follows.
5.1. RQ1. What are the AI Experts’ Perceptions of the Levels of Importance and Urgency of the Ten Issues?
AI experts in Taiwan assessed the top five issues that must be faced in preparing a society for AV to include (1) data privacy and cybersecurity, (2) regulation considerations, (3) infrastructure, (4) governance, and (5) public acceptance. This was the result of their assessment on the overall importance and urgency levels of ten issues, shown in Figure 11
. The first and second issues considered were under Regulation, the fourth was regarding the public sector’s role for general engagement, Governance, while the third and fifth were under the Wider Impact.
The above results are consistent with observations made on individual items within each issue, shown in Figure 1
, Figure 2
, Figure 3
, Figure 4
, Figure 5
, Figure 6
, Figure 7
, Figure 8
, Figure 9
and Figure 10
, based on both the proportion of numbers of item categorized into the group of being important and urgent, and on the distribution of means for items within an issue. To be specific, the top three issues have all or over two-thirds of items categorized as important and urgent, issues ranked fourth to sixth have some items, whereas issues ranked seventh to tenth have none or a few items categorized as important and urgent. Furthermore, the top five individual items across ten issues are (1) governance for AV in terms of road safety, (2) cybersecurity for driving safety, (3) regulations on liability regarding the owners of AVs, (4) regulations on insurance compensation, and (5) regulations on traffic rules and road construction. On the other hand, if we examine the top three individual items in the seven issues under Wider Impact, the results reveal: (1) the damage and cost regarding communication infrastructure, (2) transportation planning for AV resource control software applied to traffic-related matters, and (3) the economic benefits from software and related industries.
5.2. RQ2. What are the CS/EE Majors’ Perceptions of the Levels of Importance and Urgency of the Ten Issues?
CS/EE majors in Taiwan assessed the top five issues in preparing a society for AV to include (1) data privacy and cybersecurity, (2) regulation considerations, (3) public acceptance, (4) governance, and (5) infrastructure. They also assessed the overall importance and urgency levels of the ten issues. Furthermore, the top five individual items across ten issues were four cybersecurity items, plus a governance item: (1) information security on driving safety, (2) preventing AV-related antisocial activities, (3) protection mechanism for data leaking, (4) preventing the misuse and abuse of data, and (5) governance for AV in terms of road safety. On the other hand, if we examine the top individual items within seven issues under Wider Impact, the results reveal: (1) the traffic benefits of reducing accidents, (2) public acceptance of surveillance systems justifiable for AVs, and (3) public acceptance for who should own and control the data related to AVs.
In other words, across ten issues and 66 questionnaire items, CS/EE majors demonstrated a distinctive characteristic: they are most concerned with the issue of cybersecurity and data privacy, and considered both as the most important and most urgent tasks in preparing for a society for AVs. This observation was supported by (1) data privacy and cybersecurity as the top among ten issues, and four items regarding cybersecurity as the top four individual items among 66; and (2) public acceptance in third place among the ten issues, while two items under the issue of public acceptance with regard to data privacy and security were assessed as top individual items under Wider Impact. While they presented an attention-focused profile regarding the emergent issues and questions, the AI experts showed broader concerns, including issues on regulation, governance, infrastructure, transportation planning, and economy and industry, in addition to data privacy and cybersecurity.
5.3. RQ3. Among the Ten Issues and Background Variables, What are the Significant Predictors of Importance and Urgency of AV Development in Taiwan According to CS/EE Majors?
The results of two regression models, shown in Table 2
suggest that technology and innovation, gender, and infrastructure were found to be significant predictors of the overall importance level of AV development in Taiwan, all of which were positively associated. On the other hand, six variables were found to be significant predictors of the overall urgency level of AV development in Taiwan, four variables were positively associated—governance, economy and industry, technology and innovation, and infrastructure—while two variables were negatively related: data privacy and cybersecurity and regulation considerations. Gender was not found to be a significant predictor for urgency level.
The results suggest that, among CS/EE students, those who have higher regard for the potential benefits of AVs in relation to technology and innovation, such as new value, action service schemes, an increase in the proportion of electric vehicles, upgrading the traditional vehicle industry, and an overall new vision, who are male, and who have higher regard for the importance of strategies related to AV infrastructure, such as assessing the damage and cost, suitable environment for early development, and compatibility, are more likely to perceive a higher importance level of AV development in Taiwan. These three variables account for around one fourth of the variance.
In contrast, six issues were found to be significant predictors for the urgency level of AV development in Taiwan, and, altogether, accounted for a total of more than two-fifth of the variance. To be more specific, in addition to technology and innovation, as well as infrastructure, those who have higher regard for the urgency of strategies being taken related to AV governance upon various decisions, and those who have higher regard for the urgency of potential benefits of AV to the national economy and various industries, as well as those who have less concerns over the urgency of data privacy and cybersecurity, and those who have less concerns over the urgency of various dimensions regarding regulations, are more likely report a higher urgency level of AV development in Taiwan. It is interesting to note that, even though, as a group, the students were very concerned about cybersecurity and data privacy, as well as regulation considerations, on average, those who had a lower degree of concern over cybersecurity and data privacy were more likely to perceive AV development as urgent in Taiwan.
Furthermore, in both regression models for importance and urgency, technology and innovation was found to be a strong positive predictor, indicating that CS/EE majors have a higher regard for AV technology innovation and were ardent supporters of AV development in Taiwan. However, for the regression model of urgency, the other strong positive predictor was governance, revealing that CS/EE majors valued this factor as much as AV technology innovation when assessing the urgency level of AV development in Taiwan.
Overall, the findings of this study appear to provide empirical evidence to support the belief that people, in this case, even the AI experts and CS/EE majors, do not prefer “laissez-faire governance” [5
], even though the respondents did not pay much attention to the measurement item of 1.2 related to “laissez-faire or intervention” under the issue of governance, especially the students. In actuality, they have high regard for the importance and urgency of issues related to cybersecurity and data privacy, regulations, and governance, and they expect the public sector to play an active role in a joint effort with other sectors in the process of preparing a society for AVs. Furthermore, these results appear to corroborate the Ethically Aligned Design (EAD) of The Institute of Electrical and Electronics Engineers (IEEE) in stressing the principles of data agency, awareness of misuse, accountability, effectiveness, competence, and wellbeing [25
], and the Trustworthy AI Ethics Guidelines (2019) of the EU on the requirements for trustworthy AI, regarding privacy and data governance, technical safety, accountability, and societal and environmental wellbeing [26
To summarize, this study has contributed to advancing our knowledge regarding issues related to preparing a society for AV in the following dimensions.
First, extant studies have discussed AV, both its potential benefits to society and its risks, as well as challenges, and many of them provided expert opinions on policy-related recommendations. This study employed a survey instrument to investigate the perception of AI experts as well as CS/EE majors on ten issues associated with the implementation of AI in the future. We found both groups assessed the importance and urgency levels of wider social, economic, and environmental impact far behind their top concerns over cybersecurity and data privacy, regulation consideration, and governance strategies, such as information security on driving safety, governance of road safety thresholds, regulations related to the legal responsibility of owners, manufacturers, and R & D personnel in case of accidents, regulations related to insurance compensation, cybersecurity for the misuse of data, and so on, with results providing empirical evidence to support most recent studies calling for a role for the public sector in the regulation and governance of AVs’ potential risks [5
]. Furthermore, the student respondents of this study were found to be especially concerned about the issue of cybersecurity and data privacy, but, unfortunately, comparative policy overview studies revealed that diverse responses were adopted by different governments toward this issue, and, in reality, many of them are still in the process of coping with this potential risk [9
Second, regarding the importance and urgency levels of the wider impact of AVs, the findings suggested that both groups looked at these issues both from the perspective of the beneficial change they might bring about, such as economic benefits for the software industry and other related industries, and traffic benefits in terms of reducing the accident rate, and from the perspective of their potential risks, such as the possible damage and cost of the communication infrastructure. A third dimension was also addressed: actual coping strategies, questions related to how and who, such as how software would be applied in transportation planning and who should own and control the data on AVs and what the purpose would be of using these data. Most previous studies only discussed AVs on the level of issues related to them, but the results of the present study offer an initial understanding of how AI experts and CS/EE majors perceive the relative importance and urgency levels regarding specific aspects or questions that stem from these issues, so that future studies can conduct investigations based on this foundation.
Finally, the associations between background variables as well as diverse issues in relation to the perceptions of the importance level and the urgency level of AV development have rarely been explored. The findings of this study illuminate how a few wider-impact variables (technology innovation, infrastructure) can help predict the importance level of AV development, while some wider-impact variables (technology innovation, governance, economic benefits, infrastructure) as well as concern variables (cybersecurity and data privacy, regulations) can help predict the urgency level of AV development. Furthermore, regarding gender, previous studies (e.g., [22
]) have found that male students tend to have a higher positive attitude toward the overall impact of AV on mankind. The results of this study corroborate those of previous studies, as we also found that male CS/EE majors were more likely to report a higher importance level of AV development, but gender was not a significant predictor for the urgency level of AV development.
To conclude, the results of this study have the following research and policy-related implications. Following engineering and computer scientists, and researchers from transportation and other interdisciplinary fields, the academic landscape and international involvement mean that social scientists are now able to join the dialogue regarding preparing a society for AVs and can investigate the related issues from the general public’s perspective. As this is one of the first attempts to initiate a study for this purpose, this research has a research design that still needs to be improved. First, the survey instrument consists of ten issues and 66 measurement items. In future research, issues could be identified and re-categorized and measurement items deleted and added. Second, in terms of participants, this study has the research limitation of only including AI experts and CS/EE students; therefore, further study is recommended to improve the results by expanding the expert pool to involve urban and transportation engineers to understand their perspectives related to how AVs should change the way we live, including land use, mode choice, and relevant regulations. Then, the next step should be a more democratic approach, which involves civilians or the general public. Third, this study employed descriptive statistics and regression models to analyze data and conduct discussions. Other approaches are suggested to generate more in-depth observations and conclusions. Finally, the results of this study pinpointed the significance of issues related to cybersecurity and data privacy, regulations, and the governance of AVs, in addition to their wider economy and social and environmental impact, indicating a need for a more comprehensive examination of the first three issues. More importantly, since this study was mostly based in Taiwan and the context is socially and culturally bound, studies based on other jurisdictions are recommended.
For policymakers, the findings of this study seem to suggest we are still a long way from an AV-ready society. Among the issues related to preparing a society for AVs, in this case, AI experts and CS/EE majors are concerned about cybersecurity and data privacy, regulations on road safety, liability in terms of accidents, as well as insurance compensation, and governance-related decisions. It appears that, only when the most urgent issues are taken into consideration and dealt with, can socially beneficial decisions regarding the implementation of AV be the next steps. Take the issue of cybersecurity, for example: since “most governments have developed non-mandatory guidelines on cybersecurity best practices and researched to explore the implications of AVs on cybersecurity”, according to Taeihagh and Lim [9
], efforts to generate coping strategies and legal considerations are encouraged to be sped up. This would involve both data technology innovation, which major automobile brands are working on side by side with leading cloud platforms [27
], and the efforts of the public sector’ to create AV-specific legislation.
Finally, Taiwan has set up the goals of AV development by the year of 2025 to include (1) technology innovation on software and hardware for highly automated vehicles, (2) industry development for driverless electric minibuses and an The Society of Automotive Engineers (SAE) Lv5 system integration service, (3) well-established government-funded pilot testing, (4) regulations, and (5) educational investment in related fields [14
]. Since, currently, there is only one AV-related regulation in Taiwan, the Unmanned Vehicles Technology Innovative Experimentation Act, the findings from this study suggest that policymakers need to put more effort into regulation considerations and, at the same time, extend their commitment to cybersecurity and data privacy and governance strategies by identifying best practice, as observed from other countries and jurisdictions, as well as developing context-appropriate policies based on inclusive, democratic, diverse, and open discussions.