Next Article in Journal
Path Planning for Unmanned Aerial Vehicles in Dynamic Environments: A Novel Approach Using Improved A* and Grey Wolf Optimizer
Previous Article in Journal
Research on the Smooth Switching Control Strategy of Electric Vehicle Charging Stations Based on Photovoltaic–Storage–Charging Integration
Previous Article in Special Issue
Vehicle Lane Changing Game Model Based on Improved SVM Algorithm
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
WEVJ
Volume 15
Issue 11
10.3390/wevj15110530
wevj-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Tapping the Brakes: An Exploratory Survey of Consumers’ Perceptions of Autonomous Vehicles

by
George D. Shows
*,
Mathew Zothner
and
Pia A. Albinsson
Walker College of Business, Appalachian State University, Boone, NC 28607, USA
*
Author to whom correspondence should be addressed.
World Electr. Veh. J. 2024, 15(11), 530; https://doi.org/10.3390/wevj15110530
Submission received: 1 October 2024 / Revised: 29 October 2024 / Accepted: 14 November 2024 / Published: 18 November 2024
(This article belongs to the Special Issue Advancements in Autonomous Vehicles: Security, Optimization and Future Challenges)
Browse Figures
Versions Notes

Abstract

:
The purpose of this study is to gain a better understanding of the difficulty in measuring consumer acceptance of emergent technologies where artificial intelligence is present in autonomous vehicles (AVs). Using the Technology Acceptance Model (TAM) as our theoretical lens, survey data of US adult consumers are used to better understand consumer acceptance of AVs. Results from Partial Least Squares–Structural Equation Modeling (PLS-SEM) show that the certainty of product performance and interest are positively related to usage. Surprisingly, the relationship between two variables, internal locus of control and ease of use and usage, was not significant, which could be explained by AVs being self-driving and the ease of use therefore not being important in this context. Internal locus of control was negatively related to willingness to buy, and interest and usage were positively related to willingness to buy. Mediation analysis further explains these relationships. This research calls into question the TAM, long used as a measurement for the acceptance of information systems, as an acceptable model for measuring consumer acceptance where the intent is to purchase technology that contains artificial intelligence.

1. Introduction

One of the issues concerning the future of private ownership of autonomous vehicles (AVs), i.e., those that take complete control of an automobile’s direction, speed, and destination, will be consumer acceptance in the marketplace. Consumers are already purchasing and driving vehicles with some form of autonomy. Currently, the highest level of autonomy in vehicles for private consumption peaks at driver assistance, in which the vehicle controls steering and braking below 60 km, which is found in nearly one out of every four cars made today [1]. Partially assisted driving, which includes cruise control or steering, is featured in 70% of all vehicles today, leaving only 7% of all vehicles currently sold in the United States with no autonomous driving capability [1]. The leap to automated driving systems called conditional automation, however, is not commercially available in the United States. Conditional automation vehicles exist in Japan, Germany, and South Korea with restrictions, including operating in daylight, with no rain, and at limited speed ranges [2]. Fully autonomous vehicles, which take complete control of the vehicle independent of operators, are in the initial phase of testing, with approximately 1,400 self-driving vehicles being tested by over 80 companies [3].
Extant research shows contrasting but not necessarily opposing conclusions concerning their future. The Robotics Review reports both that consumer acceptance of autonomous vehicles is climbing [3] and that consumer groups state that self-driving cars are not ready for public roads [4]. The first crash of a vehicle in which a driver was charged with manslaughter while on autopilot has occurred [5]. This furthers the time to the day when AVs are accepted, as research has confirmed that fear about AVs increases in consumers with the number of accidents reported [6]. Yet, with these difficulties ahead, both automobile and trucking companies plan for a future in which AVs are accepted, desired, and considered valued assets on roads across the world.
Autonomous vehicle technology promises many benefits, including a reduction in fuel consumption [7] and a reduction in automobile fatalities through a reduction in driver fatigue [8]. For AV manufacturers to achieve these goals, however, they will first require a generation of processors with enough memory and processing power to handle all reasonable road conditions, and a database of potential sharable scenarios for AVs to “learn” from. The AV fleet company Waymo suggested that 3 million miles of test drives is not sufficient and that a small fleet of 100 vehicles would generate in excess of 1 million hours of video recording in a year [9]. Artificial intelligence (AI) holds promise in providing enough highly realistic behaviors to the AV system for the AV to house a behavioral system acceptable to handle all future possibilities [10]. It has been calculated that to train an AV fleet of 100 vehicles, a sufficient training time would be approximately 166 to 778 days and would involve up to 104 to 487 terabytes of data [11].
The promise of AVs has led to the promotion of a future that may still be a generation away [12]. As time passes without a viable commercial product, reluctance towards an AV future that realizes its potential increases as the industry grapples with the enormous technological hurdles to overcome. The AV industry has currently invested over USD 206 billion with the hope of a profitable future with no marketable product available for sale [12]. A successful sustainable future with the promise of autonomous vehicles will require a market that accepts and desires the advantages an AV future can create.
A key component of a commercially successful future in the private and public ownership of AVs will be the successful implementation of a marketing strategy that effectively answers consumer concerns and broaden awareness of the advantages of AV technology. Our goal in this paper is to investigate consumer motivations and concerns about using AVs by considering several relevant constructs tied to consumer acceptance of new technology; locus of control (LOC), certainty of product performance (COPP), ease of use (EOU), usage situations (usage), interest in the technology (interest), and the willingness for consumers to give control to the vehicle i.e., willingness to buy (WOB). Specifically, we provide a conceptual model of the relationships between these variables and consumers’ willingness to buy AVs. Our research contributes to the literature by providing a more comprehensive framework for understanding consumer motivation for and deterrence from the adoption of AVs.

1.1. Autonomous Vehicles

Autonomous vehicles are defined as vehicles that can sense their environment without human involvement [13]. The definition of “autonomy” and “involvement” accepts that there is not a single answer but permutations. The Society of Automotive Engineers (SAE) has developed an industry-accepted six-dimensional framework to describe the differing levels of automation in driving, starting at zero (meaning the driver is completely responsible for the control of the vehicle) and ending at five (in which no driver intervention is required) [14]. The ranges of automation are separated by whether a human driver monitors the driving environment, or an automated driving system monitors the driving environment. Currently, automobiles have achieved an SAE level of two, in which a human driver monitors the vehicle while the automated system controls both steering and acceleration/deceleration. The great leap forward in AVs is SAE Level 3, known as conditional automation, in which the automated system will monitor the driving environment and ask for human intervention when the time requires. SAE Level 4, high automation, will not require intervention from the human if they are unable to respond appropriately to a situation, but human intervention is still expected. Finally, SAE Level 5 is described as an autonomous system that will handle all aspects of driving under all road conditions and can manage all driving demands without the need for intervention from a human driver (see Table 1).
Autonomous vehicles have had an interesting if somewhat fractious advance in the automotive industry. Autonomous control of an automobile was introduced in the 1950s with the introduction of cruise control [15]. Autonomous controls in recent years have included advances such as adaptive cruise control (in which the speed of the vehicle is controlled by radar which detects vehicles in front), anti-lock brakes (in which the brakes do not lock up from human force, but instead reduce motion), emergency stop (where the vehicle automatically stops by detecting an object ahead), and park assist (where the car will automatically parallel park a vehicle). All of these come under Level 2 (partial automation), where the vehicle can perform steering and acceleration, but the human still monitors the tasks and can take control at any time. Levels 0 through 2 are where the human monitors the driving environment. It is Levels 3 through 5 where the automated system monitors the driving environment, effectively making decisions that the human driver accepts or overrides.
Automated systems for automobiles are examples of artificial intelligence created for a very specific purpose which requires a considerably broad understanding of driving and driving conditions [16,17]. Autonomous decision-making processes in automobiles are created to be adaptable to the constantly changing scenarios on the road. The process of creating an AI system capable of meeting SAE Level 5 is complex, involving large-scale storage of possible driving scenarios fed into a computer-designed neural network of graphic processing units (GPUs) and central processing units (CPUs) capable of handling all possible scenarios without any human interaction [10]. The complete process, from collecting the scenarios required and feeding them into an AI system capable of completing the learning process, can take upwards of a year to years [10].
When meeting these scenarios, AV decision-making may or may not reflect the decision-making processes of the driver. AV decisions can be confusing to the drivers because the vehicle could be gathering information in ways the driver is not aware of [17]. These drivers perceive the automobile acting in unanticipated ways and they experience, from their perspective, unpredictable behavior by the vehicle. This can lead to anxiety in the driver and can initiate a lack of trust in the vehicle’s capability [18].
Unpredictability in an AV can be a delimiting factor in consumer acceptance and willingness to purchase such vehicles, as technology acceptance is heavily determined by trust and the perceived risk [19,20]. According to Lee and See [20] (p. 54), trust is “the attitude, that an agent will help achieve and individual’s goals in a situation characterized by uncertainty and vulnerability.” Moving towards acceptance of AVs taking control of the tasks necessary to maneuver the passengers over distances will involve engaging in technologically and socially disruptive situations [21].
Concern over consumer safety while operating an autonomous vehicle gives rise to apprehension over whether or not the AV is operating as expected. Debate has been sparked among researchers and legislators as to the “ethical” decision-making process of an autonomous vehicle. Some researchers consider that a code of ethics will need to be written into the systems [16,17]. According to Evan Ackerman of IEEE Spectrum (the official magazine of the Institute of Electrical and Electronics Engineers), AVs will “decide” which actions to take when accidents occur: “At some point in the nearer-than-might-be-comfortable future, an autonomous vehicle will find itself in a situation where something has gone wrong, and it has two options: either it can make a maneuver that will keep its passenger safe while putting a pedestrian at risk, or it can make a different maneuver that will keep the pedestrian safe while putting its passenger at risk” [21]. Thus, the concern is whether the vehicle should protect the passengers at all costs or act to minimize the number of casualties as a whole.
While much research in AV acceptance focuses on consumer acceptance of AVs in situations in which autonomous vehicles are shared or leased for a limited period [22], the current research concentrates on a deeper commitment: a consumer’s willingness to buy an autonomous vehicle. Previous research has been performed on autonomous vehicles concerning consumer willingness to pay (WTP) and willingness to buy (WTB) regarding AVs [23]. WTB and WTP are considered unique constructs as WTB implies one is willing to buy something without consideration of price range, while WTP shows how a consumer would evaluate a product with monetary concerns in mind [23,24]. This research considers WTB to be a more valuable output of AV acceptance as it gives a more accurate reflection of a consumer’s willingness to own a vehicle when cost is not a mediating factor (see Figure 1 for our conceptual model).

1.2. Locus of Control

Locus of control (LOC) is a psychological concept that references how strongly an individual believes they have control over situations that affect their lives [25]. LOC is usually separated into two different dimensions: internal locus of control and external locus of control. Internal LOC is an individual personality trait that describes the degree to which an individual believes that they can influence the events in their lives [26]. Internal LOC entails the level to which individuals control their own behavior. External locus of control (external LOC) is where an individual believes fate, luck, or some other powerful forces control their lives [27].
In choices of technology, the relationship between locus of control and trust in technology is mixed. In an early experiment, subjects with a higher internal locus of control were found to have generally more positive attitudes towards computers than those with a higher external LOC [28]. However, in a more recent experiment with the locus of control and robotic systems, subjects with a higher external LOC were found to be negatively correlated with trust in the robot systems [28].
Because high internal LOC is related to individuals with a firm belief that they control their own destiny, they tend to look to themselves for problem-solving and have greater creativity than those with high external LOC [29]. It is posited that individuals with a high internal LOC will also not be as willing to give control to an autonomous vehicle and not be as trusting of an AV to perform its function. Thus, it is hypothesized that a high internal LOC is negatively related to willingness to buy an autonomous vehicle. A person with a high internal LOC (LOC hereafter) would perceive a limited number of benefits from a lack of trust in the AV and is likely to receive few perceived benefits from the use.
H1. 
Internal locus of control is negatively related to (H1a) usage and (H1b) willingness to buy an autonomous vehicle.

1.3. Certainty of Product Performance (COPP)

Uncertainty is a dimension of consumer risk [30]. Feelings of uncertainty may rise when the outcome of a purchase is only known after the purchase intention [31,32]. Weathers, Sharma, and Wood [30] developed the certainty of product performance (COPP) scale to assist in determining performance uncertainty for search goods and experience goods. It is used to measure the degree to which a consumer believes a product will function well and as it is intended to, especially in situations where the consumer lacks the capacity to completely experience the good through physical interaction. The search/experience distinction is based on the extent to which consumers can evaluate goods or their attributes prior to purchase [33,34]. Search goods are perceived by consumers as items that can be evaluated through the accumulation of knowledge [30]. Experience goods are classified on the extent to which consumers feel they need to directly experience goods to evaluate quality. The greater the need to use one’s senses to evaluate a good, the more experience qualities the good possesses [30].
Certainty of product performance rises as a consumer experiences the good through both pursuit of product knowledge and the experience of interacting with the product before the purchase decision. Giving consumers control over the information is effective in reducing product uncertainty [30]. COPP is further enhanced when the information is obtained from a third party, compared to obtaining it from the manufacturer or retailer. It is thus posited then that certainty of product performance is positively related to willingness to buy. Confidence in the performance of an autonomous vehicle is expected to be positively related to the willingness of a consumer to use it.
H2. 
Certainty of product performance is positively related to (H2a) usage and (H2b) willingness to buy an autonomous vehicle.

1.4. Ease of Use, Usage, and the Technology Acceptance Model

The Technology Acceptance Model (TAM) has been used as a measure of acceptance of information systems since 1986 [35]. The popularity of the TAM expanded to include no less than 698 journal citations since 2003 [35]. The TAM assumes that acceptance is predicated on two primary variables: perceived usefulness (PU) and perceived ease of use (PEOU). While initially proposed as a model for information system acceptance, its appeal has broadened to include the acceptance of autonomous vehicles. The TAM tests customer acceptance when testing autonomous vehicles on modeling software [36]. Koul and Eydgahi confirmed that perceived usefulness (PU) is the most significant predictor of future usage [32].
Ease of use (EOU) concerns the complexity and how much training is required to grasp, understand, and master a concept. Ease of use is a vital determinant in the acceptance of technology [35]. Defined as an indicator of the cognitive effort expended to utilize an innovation, ease of use is envisioned as a salient belief that affects an individual’s attitude toward the use of and mastery of a technological innovation [33]. Panagiotopoulos and Dimitrakopoulos [37] confirmed that perceived ease of use is positively related to perceived usefulness and behavioral intention to use/accept AV technology.
While EOU has been examined as a determinate of several studies that determine a consumer’s intention to accept or reject the use of new technology [38,39], Ribeiro et al. [38] contend that these studies of technology acceptance are descriptive in nature or are insufficient for explaining technology acceptance and adoption of artificial intelligence devices. Their supposition is that EOU may not be appropriate for the evaluation of artificial intelligence in AVs as consumers will be more interested in whether AVs have the same or better performance than human-driven cars [40,41], and therefore trust and perceived risk are more important. It is further noted that this experiment pertains to the willingness to buy, a commitment greater than a willingness to use a product.
To explore the viability of EOU, this research removes the concept from the Technology Acceptance Model and allows the dimension to be tested outside the construct of the TAM. Given the inconsistent history of EOU in the discussion of autonomous vehicles, this research will start from the position of supporting ease of use to be significantly related to the willingness to buy an autonomous vehicle. Usage, the potential to use an autonomous vehicle, is found in research to be positively related to an AV’s ease of use [40]. We posit that the potential use of a vehicle is positively related to a consumer’s willingness to buy.
H3. 
Ease of use is positively related to (H3a) usage and (H3b) willingness to buy an autonomous vehicle.
H4. 
Interest is positively related to (H4a) usage and (H4b) willingness to buy an autonomous vehicle.
H5. 
Usage is positively related to willingness to buy an autonomous vehicle.

2. Materials and Methods

An exploratory survey administered online via Qualtrics Survey software was conducted to measure consumers’ willingness to buy and attitudes concerning locus of control, ease of use, certainty of product performance, interest, and usage. The total survey time for respondents was approximately between ten and fifteen minutes. The study was conducted at a large public university in the southeastern United States. Undergraduate business students were requested to complete a Qualtrics Survey instrument embedded in an email link and complete the survey for the fulfillment of extra credit. Descriptive information on respondent demographics was collected. They were encouraged to share the survey with others outside the university sphere. The survey was completed by 259 respondents between the ages of 17 and 72 years old with an average age of 36.27 years. Forty-six percent were female. The sample frame included individuals of various educational backgrounds and professions. After removing incomplete surveys, a total of 233 data points were included in the succeeding analysis.
To ensure that respondents had basic knowledge levels of autonomous vehicles, after providing an informed consent form, the survey began with definitions of terms and basic information on autonomous vehicles. The information included descriptions of the driving automation levels provided by the National Highway Traffic and Safety Administration. Additionally, respondents were given a two-minute video to watch, describing and showing an autonomous vehicle driving under controlled conditions. After the video presentation, respondents were asked to begin the survey.
Cognitive responses were measured for the purpose of obtaining respondents’ spontaneous reactions to AVs and creating a Word Cloud. Subjects were first given the request to write down the first three words that came to mind when reading the phrase “Autonomous Vehicles.” Next, respondents were asked to answer questions related to autonomous vehicles using the following scales: the willingness to buy, a five-item scale [42]; internal locus of control, a three-item scale [27]; certainty of product performance, a three-item scale [29]; and ease of use, a three-item scale [29]. Interest was measured with a 3-item scale anchored by the stem, “I am interested in …”, and usage was measured with 4 items adapted from the stem, “I would use…”. The scales were 7-point Likert scales anchored with “Strongly Disagree” and “Strongly Agree”. See Appendix A for all items.

2.1. Data Analysis

Word Cloud—Cognitive Responses

Stanley Milgram [43] initiated the use of Word Clouds when he asked people to visualize landmarks in Paris. Word Clouds create a “vernacular visualization” [44] (p. 52) and are aggregators of thought, visualizing common themes in an understandable form. Two Word Clouds were compiled by using the website worditout.com. The website creates Word Clouds by having the user copy and paste the words into the website; the software compiles the word count and creates the graphic.

2.2. Quantitative Analysis

Due to the exploratory nature of the conceptual model, Partial Least Squares–Structural Equation Modeling (PLS-SEM) via SmartPLS 3.2.8 [45] was deemed more appropriate to assess the quantitative data than covariance-based Structural Equation Modeling (CB-SEM) [46]. We assessed the measurement (outer) and structural (inner) models with the structural model, providing support for H1, H4, and H5. Procedures consistent with recommendations from Hair et al. [45,47] were utilized to analyze the data and present the findings.

3. Results

The first Word Cloud was a compilation of all the cognitive responses (i.e., words described by the respondents). In this cloud, “future” was the most mentioned, with the next being “driverless”, “robot”, “futuristic”, “dangerous”, “scary” and “technology”. In this cloud, only “dangerous” and “scary” convey an emotional state (see Figure 2).
A second Word Cloud was performed using only the first cognitive response (e.g., the first word of three) given by the respondents. In this, “driverless” came first, followed by “scary”, “future”, “futuristic”, “automatic”, “interesting”, “innovative”, and “dangerous”. Notice that in the two Word Clouds, the emotions conveyed through “scary” and “dangerous” are generally fear and concern (See Figure 3).

3.1. Measurement Model Results

We assessed the psychometric properties of the scales depicted in the conceptual model Via Partial Least Squared Structural Equation Modeling (PLS-SEM) using SmartPLS 4.0 [48]. Assessing the measurement model which is referred to as Confirmatory Composite Analysis (CCA) in SmartPLS shows that most reflective indicator loadings exceeded 0.708 [49]. While one reflective loading for locus of control was 0.51, the others were acceptable for exploratory research, i.e., 0.68 and 0.91. We decided to retain the low-loading item as this research is considered exploratory and the scales are used in a new context [50]. Hair et al. [51] recommend keeping items between 0.40 and 0.070 if removing them affects content validity and if dropping them does not drastically increase the internal consistency reliability or convergent validity.
Both Cronbach’s alpha (α) and composite reliability (CR) were used to assess the reliability of all the constructs, and all contracts except for locus of control (α = 0.55) exceeded the minimum 0.70 threshold for Cronbach’s alpha. The alpha of 0.55 approaches moderate and adequate reliability (0.61–0.65) as deemed by Nunnally and Bernstein [52]. In addition, all constructs including locus of control show composite reliability above 0.60, ranging from 0.76 to 0.93, which is an indication of internal consistency reliability. As Cronbach’s alpha is a conservative measure of reliability, and composite reliability tends to overestimate the internal consistency reliability, the true reliability usually lies in between. As locus of control shows a Cronbach’s alpha of 0.55 and composite reliability of 0.76, we therefore decided to keep it in the structural model. In addition, both interest and usage were slightly over 0.90; however, as these are established scales, we decided to keep these in the model, and they were still below 0.95, which is the absolute maximum threshold recommended [51]. All constructs exceeded the 0.50 average variance extracted (AVE), ranging from 0.52 to 0.83, confirming convergent validity. When examining the collinearity statistics, two willingness to buy items had high variance inflation factors (VIFs) (>9), so these were removed, and we checked for construct reliability, construct validity, and discriminant validity again with no major changes in the results (See Table 2).
Discriminant validity was established using the Heterotrait–Monotrait (HTMT) scores, which varied between 0.29 and 0.83; these were well below 0.85 for different constructs and below 0.90 for similar constructs [51]. The HTMT score for usage and interest was 0.89, which is still below 0.9. Therefore, the measurement model was deemed to be valid and reliable. The R2 was 0.672 for usage and 0.612 for WTB in the measurement model, which shows the high explanatory power of the model’s predictive capabilities [51]. To analyze the structural model, bootstrapping was performed with 10,000 resamples. The results are reported below.

3.2. Boostrapping Results

Locus of control (LOC) is not negatively related to usage (β = 0.003, p = 0.44) but is negatively related to the WTB autonomous vehicles β = −0.13, p < 0.000; therefore, the results do not support H1a but support H1b. The second hypothesis predicting that certainty of product performance (COPP) is positively related to both usage (H2a) and WTB (H2b) only supported the path to usage (β = 0.25, p < 0.000), but not to willingness to buy (WTB) (β = 0.11, p = 0.11), thereby supporting H2a but rejecting H2b. Ease of use was neither positively related to usage β = 0.02, p = 0.71 nor to WTB, (β = 0.09, p = 0.13); thus, H3a and H3b are not supported. H4a and H4b predicted that interest is positively related to (H4a) usage and (H4b) WTB, and were both supported (β = 0.64, p < 0.000, and β = 0.59, p < 0.000, respectively). Lastly, H5 predicted a positive relationship between usage and WTB β = 0.24, p < 0.000, which was supported. While fit indices are not the norm for PLS-SEM reporting of results (in fact, Hair et al. [47,51] advise against it), the standardized root mean square residual (SRMR) was 0.07, which in CB-SEM would be considered a good fit [53]. However, as PLS-SEM and CB-SEM have different aims, this fit index should be interpreted with caution. See Table 3 and Figure 4 for the PLS-SEM results.

3.3. Ad Hoc Analysis for Mediation

Additionally, we tested whether usage mediated the paths between the independent variables and WTB. We found no mediaton for usage between ease of use (CI: 0.02–0.11, t-value 0.35, p = 0.73) and WTB or locus of control and willingness to buy (CI: 0.01–0.03, t-value 0.76, p = 0.45) as zero is within the confidence intervals (CIs). The indirect effect between COPP and WTB is significant (CI 0.018–0.113, t-value 2.526, p = 0.01). Similarly, the indirect effect between interest and WTB is significant (CI 0.051–0.252, t-value = 3.102, p = 0.002). For COPP, we confirm that usage fully mediates the relationship between COPP and WTB as the direct effect between COPP and WTB is not significant. On the other hand, for interest, usage partially mediates the relationship between interest and WTB as both the direct and the indirect effects are significant. To further substantiate the type of partial mediation, as the product of the direct and indirect effects is positive (e.g., 0.437 × 0.156 = 0.068), this suggests complementary mediation [54].

4. Discussion LOC, COPP, Interest, and Usage Were Significant

The preceding results underscore the difficulty in researching future technology that contains a fundamental shift in social and societal change. For autonomous vehicles to succeed in the marketplace, consumer acceptance by drivers and passengers to allow the control of vehicles to be given to artificial intelligence and to willingly acquiesce to the judgment of such intelligence is needed. This research attempts to pursue this deeper by investigating consumers’ willingness to purchase such a vehicle.
Many benefits pointed out when discussing the future of autonomous automobiles include a reduction in traffic deaths, a drop in harmful emissions, smoothing out of stop-and-go traffic, an improvement in fuel economy, and a reduction in travel time [55]. Yet, the future benefits will be difficult to realize if consumers are fearful of using the autonomous capabilities of vehicles. Many scenarios involving autonomous vehicles do not include the purchase of a vehicle; this includes using an AV for ride sharing, vehicle rental and leasing, in-city transportation, and taxi service. Real efficiencies of scale occur when there is a sufficient number of vehicles used by drivers in actual adoption via purchase. Furthermore, the benefits listed will not occur on a dramatic scale unless the average driver accepts them as a transportation option in their everyday use.
For these considerations, the future of autonomous vehicles requires listening to consumers and their concerns and adjusting accordingly. Consumers with a high certainty of product performance require the ability to measure the performance certainty in research on AVs as well as experience it during the search process. This will require the availability of detailed information that confirms the AV’s ability to perform in the expected manner of a vehicle that takes control of all the driving functions. It is important that the detailed information accurately reflects the AV’s qualities and not just the desirable ones, as the critical component is uncertainty, i.e., deliberately withholding information because it is detrimental to the message will have a greater impact than disclosing the information. Consumers with a high COPP will also desire a test drive of the vehicle not in a track setting, but on the streets and highways where such a vehicle would normally be used. Finally, third-party information detailing the qualities of an AV would be considered most desirable by the high-COPP individual, so reviews by consumer groups that detail the advantageous properties of an autonomous vehicle should be presented along with company information.
The fact that ease of use is not significant could lead to a reconsideration of using the Technology Acceptance Model (TAM) as a structure for considering the willingness to embrace technology when artificial intelligence (AI) is a primary component. As of this writing, there are nearly two dozen journal articles detailing some form of the TAM used in the acceptance of AVs [21,22,23,33,35,36,37,38,39,40,41,56]. This research confirms the supposition that the TAM may not be the most appropriate evaluation framework when considering AI-dominant devices. TAM articles that discuss other technologies with AI as a primary component show similar results when attempting to measure ease of use for voice assistants such as Alexa [33]. Measuring the impact of AI on the acceptance of technology may require an adjustment of the framework to be a more accurate reflection of the factors that influence a consumer’s acceptance of technology.

Moving Forward and Limitations

Future research should focus on the potential consumers of autonomous vehicles that have a high internal locus of control and certainty of product performance. Understanding the size of the audience will be critical in designing a brand message that promotes the AV to maximum benefit. Additional research needs to be performed to confirm the viability of the Technology Acceptance Model to determine whether the construct needs modification to account for AI. This research did not focus on willingness to pay, which includes the willingness to buy an item with monetary concerns in mind.
Finally, research on the acceptance of autonomous vehicles needs to include individuals who drive professionally in the United States. This includes owners/operators of large commercial trucks, drivers of passenger vehicles who act as couriers, drivers who deliver large and small packages, and drivers who ferry individuals inter- and intra-city such as Uber and taxi services. Autonomous technology does not necessarily compete for their professions any more than automatic pilots compete for airline pilot jobs; it is probable that AVs will augment the professional driver experience. However, it is also probable that professional drivers exhibit a high internal locus of control and high certainty of product performance. Research into their concerns about a future of sharing the highway with autonomous vehicles could yield valuable information, which may help guide the future acceptance of AVs. Future research could look at social media data or public forums provided by specific brands of AVs which discuss technology advancements to further investigate positive versus negative emotions through sentiment analysis [53]. In a recent study, researchers suggested that social media needs to be leveraged for hybrid and autonomous transportation as they found that only 0.6 percent of tweets in their sample of 1.25 million tweets on emerging mobility trends discussed vehicle technology.
As with any study, this research has some limitations. The study is a cross-sectional study on a topic where not many respondents have experienced the product in real life. They may have read about the product and the technology behind the autonomous steering of the vehicles but very few have adopted these vehicles yet. In addition, while the sample included non-students and non-traditional students who were older than most undergraduate students, it was a convenience sample, which may not lead to generalizable results. Future research could survey a larger, more representative sample. However, the insights presented in the current article are still valuable as they inform both academics and practitioners of consumers’ perception of AVs.

Author Contributions

Conceptualization, M.Z. and G.D.S.; methodology, M.Z. and G.D.S.; software, P.A.A.; validation, M.Z., G.D.S. and P.A.A.; formal analysis, G.D.S. and P.A.A.: Investigation, M.Z. and G.D.S.; resources, M.Z. and G.D.S.; data curation, M.Z., G.D.S. and P.A.A.; writing—original draft preparation, M.Z., G.D.S. and P.A.A.; writing—review and editing, G.D.S. and P.A.A.; visualization, M.Z. and G.D.S.; supervision, G.D.S.; project administration, M.Z. and G.D.S.; funding acquisition, Not applicable. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Ethical review and approval were waived for this study at the time of data collection due to the Code of Ethics and Conduct of Research of our University (https://researchprotections.appstate.edu/human-subjects-irb/consent-guidance-and-templates).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author(s).

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

  • Measurement Scales
  • Willingness to Buy Indicators
  • If I had the opportunity, I would consider purchasing an autonomous vehicle *. (strongly disagree to strongly agree)
  • I would seriously consider purchasing an autonomous vehicle (strongly disagree to strongly agree) *.
  • If a friend asked me, I would advise them to purchase an autonomous vehicle. (strongly disagree to strongly agree)
  • The probability that I would consider buying an autonomous vehicle is: (very unlikely to very likely)
  • My willingness to buy an autonomous vehicle is: (very improbable to very probable)
  • Locus of Control Indicators
  • When I am driving, I like to be in control. (strongly disagree to strongly agree).
  • When I am riding in a vehicle and someone else is driving, I feel uneasy (strongly disagree to strongly agree).
  • When I am driving the vehicle, I feel calm. (strongly disagree to strongly agree).
  • Interest Indicators
  • I am interested in riding in an autonomous vehicle.
  • I am interested in owning an autonomous vehicle.
  • I am interested in sharing an autonomous vehicle (taxi, Uber) for short trips.
  • Ease of Use Indicators
4.
Autonomous vehicles … will be complicated/will be simple.
5.
Autonomous vehicles … will be confusing/will be clear.
6.
Autonomous vehicles … will require a lot of training/will require no training at all.
  • Certainty of Product Performance Indicators
  • How certain are you that AVs would function properly—not certain/very certain
  • How well can you judge how AVs would function—hard to me to judge/easy to judge
  • I feel that AVs would probably—not work properly/work properly
  • Usage
  • I would use Autonomous Vehicles for: work
  • I would use Autonomous Vehicles for: entertainment
  • I would use Autonomous Vehicles for: information
  • I would use Autonomous Vehicles for: conversation.
  • * These items were removed before bootstrapping was run.

References

  1. Katharina, B. Cars Increasingly Ready for Autonomous Driving. 6 September 2024. Available online: https://www.statista.com/chart/25754/newly-registered-cars-by-autonomous-driving-level/ (accessed on 27 October 2024).
  2. Etherington, D. Over 1400 Self-Driving Vehicles Are Now in Testing by 80+ Companies Across the US. 11 June 2019. Available online: https://techcrunch.com/2019/06/11/over-1400-self-driving-vehicles-are-now-in-testing-by-80-companies-across-the-u-s/ (accessed on 27 October 2024).
  3. Staff, RBR. Robotics Business Review. 9 May 2019. Available online: https://www.roboticsbusinessreview.com/unmanned/consumer-acceptance-of-self-driving-cars-soars-study-says/ (accessed on 7 January 2022).
  4. Shaw, K. Robotics Business Review. 13 February 2019. Available online: https://www.roboticsbusinessreview.com/unmanned/consumer-group-says-self-driving-cars-arent-ready-for-public-roads/ (accessed on 8 January 2022).
  5. Gatewood, A. Tesla Driver in California First to Be Charged with Manslaughter over Fatal Autopilot Crash. 19 January 2022. Available online: https://knowtechie.com/tesla-driver-in-california-first-to-be-charged-with-manslaughter-over-fatal-autopilot-crash/ (accessed on 19 January 2022).
  6. Kareem, O. Public acceptance and perception of autonomous vehicles: A comprehensive review. AI Ethics 2021, 1, 355–387. [Google Scholar]
  7. Zewe, A. On the Road to Cleaner, Greener, and Faster Driving. 17 May 2022. Available online: https://news.mit.edu/2022/ai-autonomous-driving-idle-0517#:~:text=If%20every%20vehicle%20on%20the%20road%20is%20autonomous%2C,reduction%20in%20fuel%20or%20emissions%20is%20really%20incredible (accessed on 27 October 2024).
  8. Khattak, Z.H.; Lin, Z. Quantifying automated vehicle benefits in reducing driving stress: A simulation experiment approach. Front. Future Transp. 2023, 4, 1196629. [Google Scholar] [CrossRef]
  9. Grzywaczewski, A. Training AI for Self-Driving Vehicles: The Challenge of Scale. 9 October 2017. Available online: https://developer.nvidia.com/blog/training-self-driving-vehicles-challenge-scale/ (accessed on 27 October 2024).
  10. Voroninski, V. Interview by Popper Ben. Is This the Real Life? Training Autonomous Cars with Simulations The Stack Overflow Podcast. 11 October 2024. Available online: https://the-stack-overflow-podcast.simplecast.com/episodes/training-autonomous-cars-simulations/transcript (accessed on 27 October 2024).
  11. Leggett, D. ‘Stuck in Second Gear’—The Pathway to AVs. 18 October 2022. Available online: https://www.just-auto.com/interview/stuck-in-second-gear-the-pathway-to-avs/?cf-view (accessed on 27 October 2024).
  12. Holland-Letz, D.; Denedikt, K.; Thibaut, M. Mobility’s Future: An Investment Reality Check. 14 April 2021. Available online: https://www.mckinsey.com/industries/automotive-and-assembly/our-insights/mobilitys-future-an-investment-reality-check (accessed on 27 October 2024).
  13. Synopsis. What Is an Autonomous Car? January 2021. Available online: https://www.synopsys.com/automotive/what-is-autonomous-car.html (accessed on 15 January 2022).
  14. Mobility Insider. What Are the Levels of Automated Driving? 5 November 2020. Available online: https://www.aptiv.com/en/insights/article/what-are-the-levels-of-automated-driving (accessed on 18 January 2022).
  15. Shaout, A.; Jarrah, M.A. Cruise control technology review. Comput. Electr. Eng. 1997, 23, 259–271. [Google Scholar] [CrossRef]
  16. Wiegand, G.; Eiband, M.; Haubelt, M.; Hussmann, H. “I’d like an Explanation for That!” Exploring Reactions to Unexpected Autonomous Driving. In Proceedings of the 22nd International Conference on Human-Computer Interaction with Mobile Devices and Services, Oldenburg, Germany, 5–8 October 2020; ACM: New York, NY, USA, 2020; pp. 1–11. [Google Scholar]
  17. Koo, J.; Kwac, J.; Ju, W.; Steinert, M.; Leifer, L.; Nass, C. Why did my car just do that? Explaining semi-autonomous driving actions to improve driver understanding, trust and performance. Int. J. Interact. Des. Manuf. 2015, 9, 269–275. [Google Scholar] [CrossRef]
  18. Park, S. Multifaceted trust in tourism service robots. Ann. Tour. Res. 2020, 81, 1–12. [Google Scholar] [CrossRef]
  19. Ozturk, A.; Bilgihan, A.; Nusair, K.; Okumus, F. Wha Keeps the Mobile Hotel Booking Users Loyal? Investigating the Roles of Self-Efficacy, Compatibility, Perceived Ease of Use, and Perceived Convenience. Int. J. Inf. Manag. 2016, 36, 1350–1359. [Google Scholar] [CrossRef]
  20. Lee, J.D.; See, K.A. Trust in automation: Designing for appropriate reliance. Hum. Factors 2004, 46, 50–80. [Google Scholar] [CrossRef]
  21. Ackerman, E. People Want Driverless Cars with Utilitarian Ethics, Unless They’re a Passenger; We Want Autonomous Cars to Be as Safe for Everyone as Possible, as Long as They’re Safest for Us First. 23 June 2016. Available online: https://spectrum.ieee.org/people-want-driverless-cars-with-utilitarian-ethics-unless-theyre-a-passenger (accessed on 22 January 2022).
  22. Krueger, R.; Rashidi, T.; Rose, J.M. Preferences for Shared Autonomous Vehicles. Transp. Res. Part C Emerg. Technol. 2016, 69, 343–355. [Google Scholar] [CrossRef]
  23. Asgari, H.; Jin, X. Incorporating attitudinal factors to examine adoption of and willingness to pay for autonomous vehicles. Transp. Res. Rec. 2019, 2673, 418–429. [Google Scholar] [CrossRef]
  24. Wertenbroch, K.; Skiera, B. Measuring Consumers’ Willingness to Pay at the Point of Purchase. J. Mark. Res. 2002, 39, 228–241. [Google Scholar] [CrossRef]
  25. Glossary of Education. Locus of Control. 25 October 2013. Available online: https://www.edglossary.org/locus-of-control/ (accessed on 12 November 2021).
  26. Levenson, H. Differentiating between internality, powerful others, and chance. In Research with the Locus of Control Construct; Lefcourt, H.M., Ed.; Vol I. Assessment Methods; Academic Press: New York, NY, USA, 1981; pp. 15–63. [Google Scholar]
  27. Levenson, H. Perceived parental antecedents of internal powerful others, and chance locus of control orientations. Dev. Psychol. 1973, 9, 260. [Google Scholar] [CrossRef]
  28. Coovert, M.D.; Melvin, G. Locus of Control as a Predictor of Users’ Attitude Toward Computers. Psychol. Rep. 1980, 47, 1167–1173. [Google Scholar] [CrossRef]
  29. Glover, J.; Sautter, F. An investigation of the relationship of four components of creativity to locus of control. Soc. Behav. Personal. 1976, 4, 257–260. [Google Scholar] [CrossRef]
  30. Weathers, D.; Sharma, S.; Wood, S.L. Effects of online communication practices on consumer perceptions of performance uncertainty for search and experience goods. J. Retail. 2007, 83, 393–401. [Google Scholar] [CrossRef]
  31. Dowling, G.R.; Staelin, R. A Model of Perceived Risk and Intended Risk-Handling Activity. J. Consum. Res. 1994, 21, 119–134. [Google Scholar] [CrossRef]
  32. Havlena, W.J.; DeSarbo, W.S. On the Measurement of Percieved Consumer Risk. Decis. Sci. 1991, 22, 927–939. [Google Scholar] [CrossRef]
  33. Agarwal, R.; Prasad, J. The antecedents and consequents of user perceptions in information technology adoption. Decis. Support Syst. 1998, 22, 15–29. [Google Scholar] [CrossRef]
  34. Panagiotopoulos, I.; Dimitrakopoulos, G. An empirical investigation on consumers’ intentions towards autonomous driving. Transp. Res. Part C Emerg. Technol. 2018, 95, 773–784. [Google Scholar] [CrossRef]
  35. Lee, Y.; Kozar, K.A.; Larsen, K.R. The technology acceptance model: Past, present, and future. Commun. Assoc. Inf. Syst. 2003, 12, 50. [Google Scholar] [CrossRef]
  36. Cho, Y.; Park, J.; Park, S.; Jung, E.S. Technology Acceptance Modeling based on User Experience for Autonomous Vehicles. J. Ergon. Sociey Korea 2017, 2, 87–108. [Google Scholar]
  37. Ribeiro, M.A.; Gursoy, D.; Chi, O.H. Customer Acceptance of Autonomous Vehicles in Travel and Tourism. J. Travel Res. 2021, 61, 620–636. [Google Scholar] [CrossRef]
  38. Davis, F.D.; Bagozzi, R.P.; Warshaw, P.R. User acceptance of computer technology: A comparison of two theoretical models. Manag. Sci. 1989, 35, 982–1003. [Google Scholar] [CrossRef]
  39. Lee, J.; Lee, D.; Park, Y.; Lee, S.; Ha, T. Autonomous vehicles can be shared, but a feeling of ownership is important: Examination of the influential factors for intention to use autonomous vehicles. Transp. Res. Part C 2019, 107, 411–422. [Google Scholar] [CrossRef]
  40. Cohen, S.A.; Hopkins, D. Autonomous Vehicles and the Future of Urban Tourism. Ann. Tour. Res. 2019, 74, 33–42. [Google Scholar] [CrossRef]
  41. Kellerman, A. Automated and Autonomous Spatial Mobilities; Edward Elgar: Cheltenham, UK, 2018. [Google Scholar]
  42. Dodds, W.B.; Monroe, K.B.; Grewal, D. Effects of price, brand, and store information on buyers’ product evaluations. J. Mark. Res. 1991, 28, 307–319. [Google Scholar]
  43. Jodelet, D.; Milgram, S. Psychological maps of Paris. In Environmental Psychology: People and Their Physical Settings; Routledge: Abingdon, UK, 1976; pp. 104–124. [Google Scholar]
  44. Viégas, F.B.; Wattenberg, M. Timelines tag clouds and the case for vernacular visualization. Interactions 2008, 15, 49–52. [Google Scholar] [CrossRef]
  45. Hair, J.F., Jr. Multivariate Data Analysis; Hair, J.F., Jr., Barry, J.B., Rolph, E.A., Rolph, E.A., Eds.; Anderson Seventh Edition; Springer Science & Business Media: Berlin/Heidelberg, Germany, 2012. [Google Scholar]
  46. Ringle, C.M.; Sarstedt, M. Gain more insight from your PLS-SEM results: The importance-performance map analysis. Ind. Manag. Data Syst. 2016, 116, 1865–1886. [Google Scholar] [CrossRef]
  47. Hair, J.F.; Sarstedt Marko, M.; Christian, R.; Gudergan Siggi, S. Advanced Issues in Partial Least Squares Structural Equation Modeling; SAGE Publications, Inc.: Los Angeles, CA, USA, 2017. [Google Scholar]
  48. Hauff, S.; Richter, N.F.; Marko, S.; Ringle Christian, M. Importance and performance in PLS-SEM and NCA: Introducing the combined importance-performance map analysis (cIPMA). J. Retail. Consum. Serv. 2024, 78, 1. [Google Scholar] [CrossRef]
  49. Ringle, C.M.; Wende, S.; Becker, J. Smart PLS 4: Bönningstedt: SmartPLS. 2024. Available online: https://www.smartpls.com (accessed on 27 October 2024).
  50. Hulland, J. Use of partial least squares (PLS) in strategic management research: A review of four recent studies. Strateg. Manag. J. 1999, 20, 195–204. [Google Scholar] [CrossRef]
  51. Hair, J.F.; Hult Tomas, M.; Ringle Christian, M.; Marko, S. A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM), 3rd ed.; Sage: Los Angeles, CA, USA, 2022. [Google Scholar]
  52. Buteau, E.; Lee, J. Hey Alexa, why do we use voice assistants? The driving factors of voice assistant technology use. Commun. Res. Rep. 2021, 38, 336–345. [Google Scholar] [CrossRef]
  53. Nunnally, J.; Bernstein, I. Psychometric Theory; McGraw Hill: New York, NY, USA, 1994. [Google Scholar]
  54. Hu, L.T.; Bentler, P.M. Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychol. Methods 1998, 3, 424. [Google Scholar] [CrossRef]
  55. Koul, S.; Eydgahi, A. Utilizing technology acceptance model (TAM) for driverless car technology adoption. J. Technol. Manag. Innov. 2018, 13, 37–46. [Google Scholar] [CrossRef]
  56. Alessandrini, A.; Campagna, A.; Site, P.D.; Filippi, F.; Persia, L. Automated vehicles and the rethinking of mobility and cities. Transp. Res. Procedia 2015, 5, 145–160. [Google Scholar] [CrossRef]
Wevj 15 00530 g001
Figure 1. Proposed conceptual model.
Figure 1. Proposed conceptual model.
Wevj 15 00530 g001
Wevj 15 00530 g002
Figure 2. Word Cloud—all words.
Figure 2. Word Cloud—all words.
Wevj 15 00530 g002
Wevj 15 00530 g003
Figure 3. Word Cloud—first word.
Figure 3. Word Cloud—first word.
Wevj 15 00530 g003
Wevj 15 00530 g004
Figure 4. Bootstrapping model results.
Figure 4. Bootstrapping model results.
Wevj 15 00530 g004
Table 1. Society of Automobile Engineers International’s (SAE) standard J3016.
Table 1. Society of Automobile Engineers International’s (SAE) standard J3016.
SAE LevelNameNarrative DefinitionExecution of
Steering and
Acceleration/
Deceleration		Monitoring of Driving
Environment		Fallback
Performance of Dynamic Driving Task		System
Capability (Driving Modes)
Human Driver monitors the driving environment
0No
Automation		The full-time performance of all aspects of the dynamic driving task by the human driver, even when enhanced by warnings from intervention systemsHuman DriverHuman DriverHuman Drivern/a
1Driver
Assistance		The driving mode-specific execution of either steering or acceleration/deceleration by a driver assistance system using information about the driving environment and with the expectation that the human driver performs all remaining aspects of the dynamic driving taskHuman Driver and
System		Human DriverHuman DriverSome Driving Modes
2Partial
Automation		The driving mode-specific execution of both steering and acceleration/deceleration by one or more driver assistance systems using information about the driving environment and with the expectation that the human driver performs all remaining aspects of the dynamic driving taskSystemHuman DriverHuman DriverSome Driving Modes
Automated driving system (“system”) monitors the driving environment
3Conditional AutomationThe driving mode-specific performance of all aspects of the dynamic driving task by an automated driving system with the expectation that the human driver will respond appropriately to a request to interveneSystemSystemHuman DriverSome Driving Modes
4High
Automation		The driving mode-specific performance of all aspects of the dynamic driving task by an automated driving system, even if a human driver does not respond appropriately to a request to interveneSystemSystemSystemSome Driving Modes
5Full
Automation		The full-time performance of all aspects of the dynamic driving task by an automated driving system under all roadway and environmental conditions that can be managed by a human driverSystemSystemSystemAll Driving Modes
Source: SAE International (2014) [14].
Table 2. Construct reliability, composite reliability, and convergent validity.
Table 2. Construct reliability, composite reliability, and convergent validity.
Cronbach’s AlphaComposite ReliabilityAverage Variance Extracted (AVE)R2
COPP0.780.850.70-
EOU0.740.780.65-
INTEREST0.870.880.79-
LOC0.550.760.52-
USAGE0.900.910.770.67
WTB0.900.900.830.61
Table 3. Bootstrapping results and support of hypotheses.
Table 3. Bootstrapping results and support of hypotheses.
Hypothesisβp-Value
H1a: Locus of control–usage0.0030.44Not Supported
H1b: Locus of control–WTB−0.13<0.000Supported
H2a: COPP–usage0.25<0.000Supported
H2b: COPP–WTB0.110.11Not Supported
H3a: Ease of use–usage0.020.71Not Supported
H3b: Ease of use–WTB0.090.013Not Supported
H4a: Interest–usage0.64<0.000Supported
H4b: Interest–WTB0.59<0.000Supported
H5: Usage–WTB0.24<0.000Supported
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Shows, G.D.; Zothner, M.; Albinsson, P.A. Tapping the Brakes: An Exploratory Survey of Consumers’ Perceptions of Autonomous Vehicles. World Electr. Veh. J. 2024, 15, 530. https://doi.org/10.3390/wevj15110530

AMA Style

Shows GD, Zothner M, Albinsson PA. Tapping the Brakes: An Exploratory Survey of Consumers’ Perceptions of Autonomous Vehicles. World Electric Vehicle Journal. 2024; 15(11):530. https://doi.org/10.3390/wevj15110530

Chicago/Turabian Style

Shows, George D., Mathew Zothner, and Pia A. Albinsson. 2024. "Tapping the Brakes: An Exploratory Survey of Consumers’ Perceptions of Autonomous Vehicles" World Electric Vehicle Journal 15, no. 11: 530. https://doi.org/10.3390/wevj15110530

APA Style

Shows, G. D., Zothner, M., & Albinsson, P. A. (2024). Tapping the Brakes: An Exploratory Survey of Consumers’ Perceptions of Autonomous Vehicles. World Electric Vehicle Journal, 15(11), 530. https://doi.org/10.3390/wevj15110530

Article Metrics

Back to TopTop