An Exploration of Users’ Trust in and Willingness to Pay for Fully Driverless Vehicles After Their First Ride

Abstract

Autonomous Vehicles (AVs) have the potential to bring in social benefits. However, realizing these benefits depends on users’ attitudes toward AVs. Previous research primarily relied on surveys and simulations to explore users’ trust and willingness to pay (WTP) for fully driverless AVs, which may not fully represent real-world experiences. Thus, for the first time, we invited participants to ride driverless AVs commercially running on public roads. All participants had no prior AV riding experience. During the experiment with a 15-kilometer urban ride, participants’ trust and WTP were measured before and after the ride. Unlike previous research, participants’ driving-related experience was not found to affect their trust in and acceptance of AVs. In contrast, the first ride experience boosted participants’ trust in AVs, and high openness in personality traits was associated with higher trust. The findings suggest that offering incentives for a first ride may increase participants’ acceptance of AVs.

Keywords

trust autonomous vehicles user experience on-road study

Introduction

Autonomous Vehicles (AVs) have become practically possible with the development of driving automation technologies. The AVs have the potential to bring in social benefits, for example, increasing traffic safety, reducing carbon emissions, and increasing the efficiency of transportation systems (Massar et al., 2021; Narayanan et al., 2020; Ye & Yamamoto, 2019), if appropriate control algorithms can be adopted. However, the realization of these benefits depends on users’ acceptance of AVs, which can further be moderated by users’ trust in AVs.

Users’ or potential users’ attitudes toward AVs have been widely investigated in previous research and the major concerns of adopting AVs include cost, safety, efficiency, etc. (Q. Zhang et al., 2023). In general, though the general public has a neutral or moderate trust in and acceptance of AVs (Chng et al., 2023; Q. Zhang et al., 2023; T. Zhang et al., 2020), the characteristics of AVs, such as driving style, system transparency, and human-machine interaction design, can influence users’ trust in and acceptance of the AVs (Ekman et al., 2019; Lee et al., 2021; Luo et al., 2020; Oliveira et al., 2020). In addition, the influence of individual differences (e.g., age, gender, personality traits), cultural differences, and social influence have also been observed (Abraham et al., 2017; Kraus et al., 2021; Pop et al., 2015).

However, as AVs have not become legal to operate on public roads in most areas, the existing studies were mostly based on survey studies or driving simulator studies and users may have never experienced AVs (Colley et al., 2022; Lee et al., 2021; Morra et al., 2019; Seet et al., 2020). The few on-road studies mostly adopted Wizard-of-Oz methods, in test fields, or in AVs with safety drivers (Detjen et al., 2020; Kim et al., 2021; Strauch et al., 2019). The riding experience with these approaches can be very different from real AV riding experience, as the risks in the experiment are well-controlled and the complex traffic scenarios can hardly be recovered.

In recent years, fully driverless AVs have become available. Given that users’ trust in a system can evolve during their interactions with the system (Tolmeijer et al., 2021) and their “first impression” of AVs may affect their future adoption decisions, we are interested in how users’ trust in and acceptance of AVs (as measured by willingness to pay [WTP]) change before and after their first ride of fully driverless AVs. At the same time, as AVs are new to most people and previous experience with similar systems can moderate one’s initial trust in the new similar systems (J. Wang et al., 2023), the influence of drivers’ experience with lower levels of driving automation was also considered. Finally, as the similarity between AV driving style and drivers’ own driving style was found to affect users’ attitudes towards AVs (Lee et al., 2021), the driving style of the participants was also considered, if they owned a driver’s license.

Being different from previous studies, this is the first time real fully driverless AVs have been used in an on-road study. The outcome of this study can validate the findings in previous survey-based, simulation-based, and field studies. The results can also inform users’ concerns about adopting AVs and provide insights on how customized strategies can be used to improve the acceptance of AVs among potential customers.

Methods

Participants

We recruited 30 people to conduct the on-road study in a fully driverless taxi operated by Pony.ai in Guangzhou, China. The experiment was conducted in October to December 2023. To mitigate the impact of morning and evening peak, the hours were from 9 am to noon and from 2 pm to 5 pm. All participants were required to have no sensory impairments and had no previous experience with AVs. This study was approved by the Ethics Compliance at the Hong Kong University of Science and Technology (HREP-2023-0246).

Procedure

As shown in Figure 1, before the experiment, participants’ personality traits (measured by the Big Five personality traits; M.-C. Wang et al., 2010), gender, age, driving experience and self-reported knowledge of AVs (measured by the 10-point Likert scale) were gathered. If they reported owning a driver’s license, their driving style (measured by the Multidimensional Driving Style Inventory; Long & Ruosong, 2019) and experiences with Advanced Driver Assistance Systems (ADAS) were also collected. In total, 10 non-drivers and 20 drivers participated in the study, totaling 20 males and 10 females, between 18 and 36 (mean: 25.17, SD: 3.66).

Figure 1.

Experimental procedure.

During the experiment, the participant sat in the rear left seat of a commercially running fully driverless AV, while two experimenters sat in the front passenger seat and the rear right seat (Figure 2). The experimental route was approximately 15 kilometers on urban public roads and took around 40 min. Before and after the experiment, participants’ trust in and WTP of AVs were collected. The trust and WTP scores ranged from 1 (not at all) to 7 (extremely), calculated by averaging the values of multiple items, as shown in Tables 1 and 2. The trust questionnaire was modified from previous work (Manchon et al., 2022). The WTP questionnaire was self-designed and focused on the comparison with ride-hailing as the fully driverless AVs operate in a similar business model.

Figure 2.

The cabin environment of the commercially running driverless taxis.

Table 1.

Questionnaire of Initial and Final Trust.

Question	Initial scale (before a ride)	Final scale (after a ride)
Q1	I would feel safe in the autonomous vehicle.	I felt safe in the autonomous vehicle.
Q2	The AV provides me with more safety compared to human drivers.	The AV provided me with more safety compared to human drivers.
Q3	I would trust the decisions of an AV in most situations.	I trusted the decisions of an AV in most situations.
Q4	If the traffic conditions are complex, I will not take an AV.
Q5	If the weather conditions are terrible (e.g., fog, glare, rain), I will not take an AV.
Q6	Rather than monitoring the traffic environment, I can focus on other activities confidently.	Rather than monitoring the traffic environment, I could focus on other activities confidently.

Table 2.

Questionnaire of WTP.

Question	Scale
Q1	I am willing to use autonomous ride-hailing to get around.
Q2	I am willing to try using autonomous ride-hailing compared with ride-hailing or taxis.
Q3	Compared with ride-hailing or taxis, I am willing to use autonomous ride-hailing when it has a price advantage.
Q4	Compared with ride-hailing or taxis, I am willing to use autonomous ride-hailing even if a higher price is needed.

Dependent Variables and Statistical Models

Models of trust and WTP were used as dependent variables and two models were built for each, one for all participants and one for drivers only. For the models of all participants, ownership of a driver’s license, age, gender, personality traits, knowledge of AVs, ride (before vs. after 1st ride), and the two-way interactions of the ride with all other variables were used as independent variables; while for models of drivers, the driver’s license was replaced with the driving style and experiences with ADAS. All models were built with Proc Mixed in “SAS OnDemand for Academics” with repeated measures considered. A backward stepwise selection of models was performed based on the Bayesian information criterion (BIC) (Vrieze, 2012). Post-hoc contrasts were conducted for all significant main effects (p < .05) using Tukey’s test.

Results

Table 3 summarizes all model results.

Table 3.

Summary of Model Results.

Model	Independent variables	F-value	p
Trust (all participants)	Ride	F (1, 45.8) = 14.05	.0005*
	Gender	F (1, 54) = 3.28	.08
	Openness of personality traits	F (1, 54) = 9.41	.003*
	Conscientiousness of personality traits	F (1, 53.7) = 1.66	.2
Trust (drivers)	Ride	F (1, 36) = 3.83	.06
Trust (drivers)	Gender	F (1, 36) = 3.55	.07
WTP (all participants)	Ride	F (1, 27) = 3.10	.09
	Gender	F (1, 25) = 1.05	.3
	Ride*Gender	F (1, 27) = 5.78	.02*
	Conscientiousness of personality traits	F (1, 25) = 1.29	.3
	Extraversion of personality traits	F (1, 25) = 0.89	.4
	Agreeableness of personality traits	F (1, 25) = 0.95	.3
WTP (drivers)	Ride	F (1, 18) = 0.82	.4
	Gender	F (1, 15) = 0.14	.7
	Ride*Gender	F (1, 18) = 5.73	.03*
	ADAS	F (1, 15) = 2.10	.2
	Risky of driving style	F (1, 15) = 0.73	.4
	Concerned of driving style	F (1, 15) = 0.83	.4

Note. *marks significant effects (p < .05).

Trust Models

We found that the Ride and the Openness of Personality Traits can influence participants’ trust in AVs. Specifically, as shown in Figures 3 and 4, a 1-unit increase in openness led to 0.38 units increase in trust (95% confidence interval [95%CI]: [0.13, 0.63]), and taking the first ride increased participants’ trust in AVs (difference (∆) = 0.74, 95%CI: [0.34, 1.13], t (45.8) = 3.75, p = .0005).

Figure 3.

The effect of openness on trust (all participants). The shadow represents the 95% confidence interval.

Figure 4.

The effect of first ride experience on trust (all participants).

WTP Models

At the same time, a significant interaction effect between Ride*Gender has been observed among all participants (Figure 5). After the first ride, male participants’ WTP increased (∆ = 0.69, 95%CI: [0.23, 1.15], t (27) = 4.12, p = .002). As for participants who owned driver’s licenses (Figure 6), only male participants’ WTP increased by 0.67 (95%CI: [0.10, 1.24], t (18) = 3.30, p = .02).

Figure 5.

WTP across rides and gender (all participants).

Figure 6.

WTP across rides and gender (drivers).

Discussions

For the first time, we validated the impact of individual differences (including, personality traits and gender) on users’ trust in and acceptance of AVs in commercially running driverless taxis. Being different from previous research (Ekman et al., 2019; Lee et al., 2021), we found that neither ownership of a driver’s license nor experiences of ADAS can affect users’ trust in and acceptance of AVs. It is possible that users cared more about driving safety instead of the match of driving style, given that users were exposed to higher risk compared to those in driving simulators or closed tracks. This indicates that for high-risk systems, simulation or survey may not fully reveal users’ attitudes toward the system. Future research with a larger sample size is still needed to validate this hypothesis.

At the same time, we found that the first ride experience can boost users’ trust in and acceptance of AVs. Most previous survey studies show that users still have concerns over AVs (Q. Zhang et al., 2023; T. Zhang et al., 2020). Our study found that inviting users to experience AVs may increase their acceptance of AVs and future AV companies may take strategies such as discounted or free first ride to attract more users, especially in the early stage of the commercial operation. The strategy, at the same time, should consider individual differences, as gender and levels of openness affected participants’ trust in and WTP for AVs, which can further be moderated by one’s driving experience.

However, it should be noted that, in all rides in our experiment, the AVs ran smoothly without failures, and still, participants were accompanied by two experimenters. It is still interesting to investigate how real on-road failures may affect users’ trust in and acceptance of AVs and when participants took a ride of the AVs alone.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (No. 52202425), the Guangzhou Municipal Science and Technology Project (No. 2023A03J0011), and the Project of Hetao Shenzhen-Hong Kong Science and Technology Innovation Cooperation Zone (HZQB-KCZYB-2020083). We gratefully acknowledge Pony.ai for their help in the experiment.

ORCID iD

Dengbo He

References

Abraham

Lee

Brady

Fitzgerald

Mehler

Reimer

Coughlin

J. F.

(2017). Autonomous vehicles and alternatives to driving: Trust, preferences, and effects of age. Proceedings of the transportation research board 96th annual meeting, Washington, DC.

Chng

Plananska

Cheah

(2023). Comparison of travel attitude study methods using online tools: The case of understanding public acceptance of autonomous vehicles. Transportation Research Interdisciplinary Perspectives, 20, 100847. https://doi.org/10.1016/j.trip.2023.100847

Colley

Rädler

Glimmann

Rukzio

(2022). Effects of scene detection, scene prediction, and maneuver planning visualizations on trust, situation awareness, and cognitive load in highly automated vehicles. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 6(2), 1–21. https://doi.org/10.1145/3534609

Detjen

Pfleging

Schneegass

(2020). A wizard of oz field study to understand non-driving-related activities, trust, and acceptance of automated vehicles. AutomotiveUI 20: 12th international conference on automotive user interfaces and interactive vehicular applications, DC, Virtual Event, USA.

Ekman

Johansson

Bligård

L.-O.

Karlsson

Strömberg

(2019). Exploring automated vehicle driving styles as a source of trust information. Transportation Research Part F: Traffic Psychology and Behaviour, 65, 268–279. https://doi.org/10.1016/j.trf.2019.07.026

Kim

Park

Lim

Y.-k.

(2021). Guiding preferred driving style using voice in autonomous vehicles: An on-road wizard-of-oz study. Proceedings of the 2021 ACM designing interactive systems conference, Virtual Event, USA.

Kraus

Scholz

Baumann

(2021). What’s driving Me? Exploration and validation of a hierarchical personality model for trust in automated driving. Human Factors: The Journal of the Human Factors and Ergonomics Society, 63(6), 1076–1105. https://doi.org/10.1177/0018720820922653

Lee

J. D.

Liu

S.-Y.

Domeyer

DinparastDjadid

(2021). Assessing drivers’ trust of automated vehicle driving styles with a two-part mixed model of intervention tendency and magnitude. Human Factors: The Journal of the Human Factors and Ergonomics Society, 63(2), 197–209. https://doi.org/10.1177/0018720819880363

Long

Ruosong

(2019). Reliability and validity of the multidimensional driving style inventory in Chinese drivers. Traffic Injury Prevention, 20(2), 152–157. https://doi.org/10.1080/15389588.2018.1542140

10.

Luo

Chu

Yang

X. J.

(2020). Trust dynamics in human-AV (automated vehicle) interaction. Extended abstracts of the 2020 CHI conference on human factors in computing systems, New York, NY, United States.

11.

Manchon

J. B.

Bueno

Navarro

(2022). How the initial level of trust in automated driving impacts drivers’ behaviour and early trust construction. Transportation Research Part F: Traffic Psychology and Behaviour, 86, 281–295. https://doi.org/10.1016/j.trf.2022.02.006

12.

Massar

Reza

Rahman

S. M.

Abdullah

S. M. H.

Jamal

Al-Ismail

F. S.

(2021). Impacts of autonomous vehicles on greenhouse gas emissions—Positive or negative? International Journal of Environmental Research and Public Health, 18(11), 5567. https://doi.org/10.3390/ijerph18115567

13.

Morra

Lamberti

Pratticó

F. G.

La Rosa

Montuschi

(2019). Building trust in autonomous vehicles: Role of virtual reality driving simulators in HMI design. IEEE Transactions on Vehicular Technology, 68(10), 9438–9450. https://doi.org/10.1109/TVT.2019.2933601

14.

Narayanan

Chaniotakis

Antoniou

. (2020). Factors affecting traffic flow efficiency implications of connected and autonomous vehicles: A review and policy recommendations. In Milakis

Thomopoulos

van Wee

(Eds.), Advances in Transport Policy and Planning (Vol. 5, pp. 1–50). Academic Press. https://doi.org/10.1016/bs.atpp.2020.02.004

15.

Oliveira

Burns

Luton

Iyer

Birrell

(2020). The influence of system transparency on trust: Evaluating interfaces in a highly automated vehicle. Transportation Research Part F: Traffic Psychology and Behaviour, 72, 280–296. https://doi.org/10.1016/j.trf.2020.06.001

16.

Pop

V. L.

Shrewsbury

Durso

F. T.

(2015). Individual differences in the calibration of trust in automation. Human Factors: The Journal of the Human Factors and Ergonomics Society, 57(4), 545–556. https://doi.org/10.1177/0018720814564422

17.

Seet

Harvy

Bose

Dragomir

Bezerianos

Thakor

(2020). Differential impact of autonomous vehicle malfunctions on human trust. IEEE Transactions on Intelligent Transportation Systems, 23(1), 548–557. https://doi.org/10.1109/TITS.2020.3013278

18.

Strauch

Mühl

Patro

Grabmaier

Reithinger

Baumann

Huckauf

(2019). Real autonomous driving from a passenger’s perspective: Two experimental investigations using gaze behaviour and trust ratings in field and simulator. Transportation Research Part F: Traffic Psychology and Behaviour, 66, 15–28. https://doi.org/10.1016/j.trf.2019.08.013

19.

Tolmeijer

Gadiraju

Ghantasala

Gupta

Bernstein

(2021). Second chance for a first impression? Trust development in intelligent system interaction. Proceedings of the 29th ACM conference on user modeling, adaptation and personalization, Utrecht, Netherlands.

20.

Vrieze

S. I.

(2012). Model selection and psychological theory: A discussion of the differences between the Akaike information criterion (AIC) and the Bayesian information criterion (BIC). Psychological Methods, 17(2), 228. https://doi.org/10.1037/a0027127

21.

Wang

Huang

(2023). Influential factors of users’ trust in the range estimation systems of battery electric vehicles—A Survey Study in China. arXiv preprint arXiv:2301.10076.

22.

Wang

M.-C.

Dai

X.-Y.

Yao

S.-Q.

(2010). Development of Chinese big five personality inventory (CBF-PI)—Theoretical framework and relability analysis. Chinese Journal of Clinical Psychology, 18(5), 545–548. https://doi.org/10.16128/j.cnki.1005-3611.2011.04.004

23.

Yamamoto

(2019). Evaluating the impact of connected and autonomous vehicles on traffic safety. Physica A: Statistical Mechanics and its Applications, 526, 121009. https://doi.org/10.1016/j.physa.2019.04.245

24.

Zhang

(2023). Human acceptance of autonomous vehicles: Research status and prospects. International Journal of Industrial Ergonomics, 95, 103458. https://doi.org/10.1016/j.ergon.2023.103458

25.

Zhang

Tao

Zhang

Zeng

Zhu

(2020). Automated vehicle acceptance in China: Social influence and initial trust are key determinants. Transportation Research Part C: Emerging Technologies, 112, 220–233. https://doi.org/10.1016/j.trc.2020.01.027