Public concerns and connected and automated vehicles: safety, privacy, and data security

Connected and automated vehicles (CAVs) are rapidly developing, raising significant societal implications. Public opinion plays a crucial role in shaping the pace and parameters of CAV adoption and regulation. While safety is a primary concern due to publicized accidents, other issues like privacy and data security are also emerging as important considerations. This study investigates the multifaceted public concerns surrounding CAVs, exploring how different demographic groups relate to specific policy issues such as safety, privacy, and data security. The researchers define 'politics' as the exercise of power in social fields, encompassing formal rules (laws) and the shaping of cognitive categories and practices. The study specifically analyzes safety (technological reliability and human elements), privacy (data collection about personal trips), and data security (risks of hacking). The existing literature emphasizes safety concerns, but this study expands the analysis to include privacy and security, acknowledging that public concerns are a crucial element in policy debates and rule-making processes. The study also explores how public opinion varies across demographic groups, such as gender, age, race, ethnicity, income, and political ideology, expecting a nuanced, rather than uniform, public opinion. The concept of 'levels of automation', as defined by SAE International, is considered, acknowledging the varying levels of public understanding and the implications for policy.

Existing research on public perceptions of CAVs primarily focuses on demographic variables like gender, age, and income as predictors of attitudes towards CAVs. Studies consistently show that men are more likely than women to be interested in and willing to accept CAVs, possibly due to higher technological savviness among men. The relationship between age and attitude towards CAVs is less consistent, with some studies showing younger individuals being more open to adoption. Income and education levels are also often correlated, with higher earners and those with higher education generally showing more familiarity and positive opinions of CAVs. Regarding race and ethnicity, some studies find no significant relationship between race and attitudes, but others reveal that Hispanics consistently hold more positive views than non-Hispanics. The relationship between political views and public acceptance or concern is less clear, although some studies suggest conservatives might be more concerned about privacy. The literature reveals a gap in comprehensive analysis of the relationships between demographic differences and concerns about safety, privacy, and data security specifically. While safety is a dominant theme, with some studies showing positive views on increased safety and convenience while others reveal negative attitudes and concern about accidents, there is a lack of systematic analysis of how demographic factors influence perceptions of safety. Similarly, there is limited research on public opinion and privacy for CAVs, with studies suggesting that the level of concern varies across countries. Finally, while data security concerns have been identified, research on demographic factors affecting those concerns is sparse. The study addresses this research gap by systematically examining the relationships between demographic variables and concerns about safety, privacy, and data security.

The study used a survey distributed through Amazon Mechanical Turk (MTurk) to collect data on public opinion. The survey included questions measuring public opinion on safety, privacy, and data security, using ordinal scales (e.g., 'very safe' to 'very unsafe'). The sample consisted of 1000 US residents aged 21 and older who were familiar with self-driving cars. After data cleaning, which involved removing incomplete responses, duplicates, and unreliable responses (based on consistency checks), the final sample size was 914. The researchers used several demographic variables as independent variables, including gender, age (grouped into categories), race (White/non-White), ethnicity (Hispanic/non-Hispanic), education (grouped into categories), income (grouped into categories), political views (conservative, moderate, liberal), and urban/rural residence (urban, suburban, rural). Proportional odds ordered logistic regression was employed to analyze the data, given the ordinal nature of the dependent variables. Bayesian multiple imputation was used to handle missing data, under the assumption of data being missing at random (MAR). The researchers acknowledge potential sampling biases associated with using MTurk (such as the potential for a sample skewed towards college-educated respondents) and address this by including educational attainment in the analysis, serving as both a variable of interest and a control variable. The researchers also note that the descriptive and univariate statistics are not generalizable to the wider public due to these sampling biases.

The proportional odds ordered logistic regression analyses revealed several significant findings: **Safety:** * Women were 1.4 times more likely than men to be concerned about CAV safety. * Respondents over 60 were 2.2 times more likely than those aged 21-30 to express safety concerns. * Non-White respondents were 1.4 times more likely than White respondents to be concerned about safety. * Hispanics were 54.8% less likely than non-Hispanics to express safety concerns. * Rural and suburban residents were 1.5 times more likely than urban residents to express safety concerns. **Privacy:** * Hispanics were 1.8 times more likely than non-Hispanics to be concerned about CAV privacy. * Liberals and Moderates were 45.4% and 38.3% less likely than Conservatives to be concerned about privacy, respectively. **Data Security:** * Non-White respondents were 1.7 times more likely than White respondents to be concerned about data security. * Those earning over $100,000 were 48.3% less likely than those earning less than $50,000 to express security concerns. * Liberals were 44.7% less likely, and Moderates were 35.4% less likely than Conservatives to express security concerns. The study notes that while some variables show consistent relationships across different public concerns (e.g., race and ethnicity, political ideology), the relationships between other demographic variables and the level of concern vary across safety, privacy, and security.

The findings highlight the diverse nature of public concerns regarding CAVs and their significant association with demographic factors. The increased concern for safety among women, older individuals, and non-White respondents suggests a need for targeted communication and policy strategies to address these specific concerns. The lower concern for privacy among liberals and moderates, and the higher concern among Hispanics, point to the complex interplay between political ideology, cultural background, and perceptions of privacy. The significant association between race and data security concerns underscores the need for addressing potential digital inequities and ensuring that marginalized groups are not disproportionately impacted by security breaches. The study emphasizes the need for granular, demographically informed policy-making in the development and deployment of CAVs, moving beyond a generalized approach to public opinion. The findings also show how different demographic groups might form coalitions around specific concerns in the ongoing political debate about CAVs.

This study provides valuable insights into the diverse public concerns surrounding CAVs, revealing significant associations with various demographic factors. The findings underscore the need for a nuanced approach to CAV policy, addressing specific concerns of different demographic groups. Future research could explore these relationships in other countries and investigate other societal concerns related to CAVs, such as governance, sustainability, and equity. Additionally, longitudinal studies are needed to track changes in public opinion as CAV technology evolves and becomes more widespread.

The study's reliance on a US-based sample limits its generalizability to other regions. The use of MTurk might have introduced some sampling biases, although the researchers attempted to mitigate this by including education as a control variable. The study's focus on three specific concerns (safety, privacy, security) might not fully capture the breadth of public concerns related to CAVs. Furthermore, the cross-sectional nature of the data limits the ability to establish causal relationships between demographic factors and attitudes.