Understanding Intention to Use Conditionally Automated Vehicles in Thailand, Based on an Extended Technology Acceptance Model

Abstract

Automated vehicles (AVs) provide several advantages in solving issues of road traffic; including enhanced safety, reduced greenhouse gas emissions, and reduced traffic congestion. As AVs are still relatively new developments in developing countries, AV adoption faces challenges from both technological and psychological issues. Therefore, our initial research focus is on identifying the factors that influence the intention to use conditionally automated vehicles (CAVs; SAE Level 3). An extended technology acceptance model (TAM), which includes Trust, Perceived Risks, and Environmental concerns, is proposed as the predictor model in this study. The 299 participants gathered through online surveys in Thailand were examined using the Structural Equation Model (SEM) technique. In this study, Trust was shown to be the strongest predictor of Intention, followed by Perceived Ease of Use, whereas Perceived Usefulness had no impact on intention to use the SAE Level 3. The results of this study will be able to guide the forming of future policies that aim at promoting the use of AVs and helping technology developers create systems to better meet the needs of users in developing nations.

1. Introduction

Automated/self-driving vehicles (AVs) have been expected to be the future of private and public transportation. AVs provide several advantages, including enhanced safety, reduced greenhouse gas emissions, and reduced traffic congestion. According to data on road safety, human error is to blame for 94 percent of car collisions [1]. AVs, therefore, have the potential to make roads safer by minimizing the need for human involvement in the driving task. AVs are supposed to be electric, which can help to reduce carbon dioxide emissions and lessen global warming because they are battery-powered. For this reason, when compared to conventional automobiles, driving battery electric vehicles (BEVs) can reduce greenhouse gas emissions by up to 64%, according to research from the University of Michigan [2].

The benefits of AVs are anticipated to be much greater in developing countries where transportation-related problems are more severe. For example, Thailand suffers from serious issues of traffic safety and greenhouse gas emissions from the road transportation. With reference to the World Health Organization (WHO) report published in 2018, Thailand had the highest road fatality rate in Southeast Asia, with 32.7 fatalities per 100,000 population [3]. A Thai finding indicated that the road areas (urban and rural) of Thai roadways had significant impact on the severity of rear-end collisions due to various factors, such as the involvement of elderly drivers, etc. [4]. Furthermore, in single-vehicle crashes, drivers who exceeded the speed limit or fatigued drivers were more likely to be involved in severe/fatal injuries [5]. Consequently, AVs could reduce these severe accidents by driving autonomously. In addition, as reported by Thailand Greenhouse Gas Management Organization, road transportation consumes the most energy and emits the greatest greenhouse gas emissions amid the entire transportation sector [6].

Promoting the use of AVs requires an understanding of user preferences and barriers. Based on international public opinion polls, the majority of people feel that the benefits of AVs are reduced fuel consumption, reduced severe accidents, and lowered insurance rates. They believed that self-driving cars would be more user-friendly than traditional vehicles, and they displayed a favorable attitude toward this self-driving technology [7,8]. However, some people were unsure of this new technology, for instance, system failure, legal liability, lack of driver control, software hacking, and so on [7,8,9]. These concerns differ depending on the degree of automation.

According to the Society of Automotive Engineers (SAE), autonomous driving technology is categorized into six levels (from 0 to 5). At levels 0 to 2, the driver is still required to operate the vehicle. SAE Level 3 is the entry-level autonomous driving, SAE Level 3 of automated driving is defined as “The entry-level autonomous driving system controls the vehicle’s movement, braking, and responding to its surroundings, but drivers must be ready to operate the vehicle when the system requires it” [10]. At level 3 and higher, the vehicle can be operated autonomously with varied limitations.

Promoting the use of AVs could help eliminate elevated traffic-related problems in developing countries. Understanding the underlying factors that affect the intention to use them is therefore necessary. However, as can be seen from the literature reviews, there is a gap in knowledge relevant to developing countries. Most recent studies were conducted in and focused on developed countries (such as the United States, China, Germany, and Korea). It could be argued that these nations are more able to adopt AVs since they have greater resources available in a variety of areas [11]. With the different policies and regulations, technological breakthroughs, infrastructures, and consumer acceptability, the information gained cannot directly be transferred to and applied in the developing countries. In addition, in a developing country, SAE Level 5 may take a great deal of time to break into the domestic market. As a result, it is more feasible to see SAE Level 3 on roads.

This study, therefore, aims to reveal psychological factors related to the intention of adopting SAE Level 3 in developing cities. It is hoped that it will provide the necessary information that is particular to the region when marketing AVs takes its place in developing nations.

2. Literature Reviews

2.1. The Technology Acceptance Model (TAM)

The technology acceptance model (TAM) is the most widely applied psychological concept in the field of technology acceptance, particularly in predicting the adoption of AVs. TAM relationships are proposed by Venkatesh and Davis (1989), who stated that usage behavior was determined by the intention to use a particular system. Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) are the two factors that determine users’ intention to use, and PEOU affects PU [12]. Regarding intention to adopt AVs, PU is defined as “the extent to which an individual believes that using AVs would improve productivity” and PEOU is defined as “the degree to which an individual believes using AVs would be effortless” [12].

Based on this concept, numerous studies have expanded the TAM model to include additional factors (such as Trust, Perceived Risk (PR), and Environmental Concern) to predict the intention to use/accept AVs more accurately.

2.2. Psychological Factors Affecting Adoption of AVs

Several studies have been conducted to investigate the psychological factors that affect people’s intentions to use Avs, as shown in

Table 1. An overview of AV research on the concept of TAM theory.

Author	Country	Automation Level	Samples	Analysis Method	Main Findings	R2 of Intention
Choi and Ji, 2015 [13]	Korea	Level 3	552	SEM	TR, PU, PEOU, ELOC→INT (+) TR→PU (+) TR→PR (−) ST, TC, SM→TR (+)	0.68
Buckley et al., 2018 [14]	U.S.	Level 3	74	Hierarchical regression	PU+PEOU+TR→INT (+)	0.44
Xu et al., 2018 [15]	China	Level 3, Level 5	300	SEM	TR, PU, PEOU, PS→INT (+) TR→PU, PEOU, PS (+)	0.40, 0.55
Zhang et al., 2019 [16]	China	Level 3	216	SEM	ATT, PU→INT (+) PEOU→PU (+) PEOU, TR→ATT (+) PU→TR (+) PSR→TR (−)	0.61
Koul and Eydgahi, 2018 [17]	U.S.	Level 5	377	Multiple Regression	PU, PEOU→INT (+) Driving Experience→INT (−)	0.62
Rahman et al., 2019 [18]	U.S.	Level 5	173	Regression	ATT+PU+TR→INT (+)	0.77
Motamedi et al., 2020 [19]	U.S.	Level 5	310	SEM	CO→TR (+) TR→PS (+) PS, PEOU→PU (+) PS, PU→INT (+)	0.91
Panagiotopoulos and Dimitrakopoulos, 2018 [20]	Greece	Level 5	483	SEM	PU, PEOU, TR, SI→INT (+) PEOU→PU (+)	0.44
Wu et al., 2019 [21]	China	Level 4 and Level 5 (implied)	470	SEM	GPU, PEOU, EC→INT (+) PEOU, EC→GPU (+) EC→PEOU (+)	N/A
Nastjuk et al., 2020 [22]	Germany	Level 5	316	SEM	ATT, TR, CO, PE→INT (+) PEOU, RA, CO→ATT (+) SN, PI, RA, CO→PU (+) PI, EN→PEOU (+)	N/A
Hein et al., 2018 [23]	Germany	Level 4	642	SEM	PU, SI→ (+) PEOU→PU (+) Technology Risk→PU (−) Work, Reading, Internal/External socialization→PU (+)	0.67
Lee et al., 2019 [24]	Korea	Level 5	313	SEM	SE, PO, PU→INT (+) PR→INT (+) RA, PEOU→PU (+) SE→PEOU (+)	0.52

Note: INT = Intention; ATT = Attitude; PU = Perceived Usefulness; PEOU = Perceived Ease of Use; TR = Trust; CO = Compatibility; PE = Price Evaluation; RA = Relative Advantage; SN = Subjective Norms; PI = Personal Innovativeness; EN = Enjoyment; PS = Perceived Safety; GPU = Green Perceived Usefulness; EC = Environmental Concern; PSR = Perceived Safety Risk; SE = Self-Efficacy; PO = Psychological Ownership; PR = Perceived Risk; SI = Social Influence; ELOC = External Locus of Control; ST = System Transparency; TC = Technical Competence; SM = Situation Management; (+) represents a positive effect; (−) represents a negative effect.

. According to the reviews, psychological elements that influence people’s intentions to use AVs vary depending on the region, level of autonomy, sociodemographics, experience with technology, etc.

Regarding SAE Level 3, Choi and Ji (2015) surveyed Koreans in order to investigate what factors affected the adoption. The results of the extended TAM model showed that PU and Trust were the most crucial variables. Trust had a negative impact on PR, and the model accounted for 68% of the variation in intention [13]. Buckley et al. (2018) conducted a 20-minute experiment using simulated driving and found that PU and Trust were significantly important to intention to use SAE Level 3. The model could explain 44% of the variance in intention [14]. The field experiment method was used in a Chinese study by Xu et al. (2018) to investigate the factors that influenced passenger acceptance. Their study discovered that factors such as trust, PU, and perceived safety influenced willingness to re-ride (WTR), with the model explaining 40% of the variance in WTR [15]. Finally, the study by Zhang et al. (2019) examined the role of Trust and PR factors (including safety risk, and privacy risk). They discovered that Trust had strong impact on attitudes toward SAE Level 3, with PU strengthening Trust, while 61% of the variance in intention could be explained by their model [16]. In summary, the critical predictors of intention to use SAE Level 3 were mostly PU and Trust, whereas PEOU showed very little or no influence on intention [13,14,15,16]. As mentioned by Xu et al. (2018), the PEOU factor had a greater influence on the intention to use SAE Level 5 than SAE Level 3 due to users being just passengers [15].

Focusing on the SAE Level 5, Koul and Eydgahi (2018) used TAM model to investigate the acceptance of driverless cars in the United States using an online questionnaire. The study revealed that PU and PEOU influenced the intention to use a driverless car [17]. Additionally, driving experience was shown to be significantly influenced as well. M.M. Rahman et al. (2019) conducted a questionnaire survey to investigate the elderly’s perceptions of SAE Level 5. According to the survey results, the model explained 77% of the variance in self-driving vehicles acceptance, and it was discovered that attitudes, PU, and Trust affected the adoption of the cars [18]. Motamedi et al. (2020) studied AVs acceptance (personally owned) by collecting data from focus groups and questionnaires. The findings demonstrated that PU, Trust, and Compatibility all appeared to play an important role in SAE Level 5 usage intentions, with the model explaining 91% of the variance in intention [19]. Further research from Greece by Panagiotopoulos and Dimitrakopoulos (2018) investigated the intention to adopt automation technology, regarding Trust and Social Influence (SI) factors in TAM, using a questionnaire. The results indicated that PU had strong impact on intention to use, followed by Trust, SI, and PEOU, respectively [20]. J. Wu et al. (2019) conducted a study in China to explore the Environmental Concerns (EC) that influenced the adoption of electric AVs using an online survey. According to the results, TAM and EC factors influenced intention to use electric AVs, and EC indirectly influenced intention as well [21]. Another study conducted by I. Nastjuk et al. (2020), numerous factors influenced the adoption of full AVs. The findings revealed that the relationship between TAM factors and Trust had a significant effect (PEOU had no influence on PU) on SAE Level 5 usage intention. Additionally, system characteristics (including relative advantage, compatibility (CO), enjoyment, and price evaluation) were found to have an impact on both PU and PEOU [22]. In conclusion, ATT, PU, and Trust strongly represented the predictors of SAE Level 5 adoption; however, other factors may play a role depending on each study (e.g., EC, PR, SI, CO) [21,22,23,24].

2.3. Research Framework

With reference to TAM relationships examined by Venkatesh and Davis (1989), the usage of AVs depends on its benefits, and the system must be easy to use and should not need too much mental effort. In addition, other factors related to individual differences are also significant in regard to intention, e.g., Trust, Perceived Risk, and Environmental Concern.

Trust is the factor that influences whether or not an individual intends to use automation technology, especially the AVs. According to Xu et al. (2018), Trust had a strong influence on PU and PEOU [15]. Furthermore, a study by Choi and Ji (2015) indicated that trust influenced PU. Trust is also a very important barrier to the adoption of AVs among older adults [25]. Therefore, the intention to use AVs depends on personal beliefs concerning technological reliability and that the system would perform well for everyday commuting purposes [13].

The Perceived Risk (PR) factor is considered alongside Trust, from the context of e-services to the acceptance of AVs. This was confirmed by Choi and Ji (2015), who found that PR had a negative effect on intention to use SAE Level 3. Thus, technical issues, system hacking, or personal data, as well as legal liability, have a negative effect on the intention toward the use of AVs in the future [13].

Environmental Concern (EC) is a factor that would play a role in adopting AVs in the future due to the causes of global warming. We can cooperate to decrease global warming by driving cars that generate no CO₂. A study by J. Wu et al. (2019) found that Environmental Concerns influenced electric AVs (SAE Level 4–5 (implied)) adoption as AVs are a low-carbon travel mode of transportation that are able to reduce air pollution [21].

Therefore, to better understand users’ intentions, the TAM theory was extended in this study by including Trust, Perceived Risk (PR), and Environmental Concern (EC) constructs to examine the intention of using Avs, as shown in Figure 1.

According to the previous research, the following hypotheses were proposed:

H1. 

Perceived usefulness positively affects the intention to use SAE Level 3.

H2. 

Perceived ease of use positively affects the intention to use SAE Level 3.

H3. 

Perceived ease of use positively affects the perceived usefulness of SAE Level 3.

H4. 

Trust positively affects the intention to use SAE Level 3.

H5. 

Trust positively affects the perceived usefulness of SAE Level 3.

H6. 

Trust positively affects the perceived ease of use of SAE Level 3.

H7. 

Perceived risk negatively affects the intention to use SAE Level 3.

H8. 

Environmental concern positively affects the intention to use SAE Level 3.

H9. 

Environmental concern positively affects the perceived usefulness of SAE Level 3.

3. Methodology

3.1. Procedure

The data were collected through a survey using an online questionnaire via Google Forms. A link was posted to Facebook groups where readers were asked to share the questionnaire with their friends and relatives (the Snowball Sampling Method). The surveyed data were then gathered via the Fastwork website.

The questionnaire was divided into four sections: general information about the respondents (4 items), driving information (5 items), psychological factors based on the extended TAM (24 items), and beliefs about AVs (1 item). Each respondent spent an average of 7 min when completing the questionnaire.

The survey focused on the SAE Level 3. Before conducting the questionnaire, SAE Level 3 had been clarified with a description and a video showcasing the use of SAE Level 3 to assist each respondent to better understand the concept.

Before distributing the online questionnaire, the study had been approved by the Human Ethics Committee with assurance that participant information was to be maintained confidential.

3.2. Participants

All the data were obtained from a total of 332 Thai respondents. Respondents gave incomplete data in more than 10% of the questions, and those with incomplete answers were eliminated. The number of samples left was 299 after data cleaning.

The respondents were 55% males and 45% females. Their ages ranged from 18 to 64 years, with the age group 21–29 years accounting for 31% and the age group 30–39 years accounting for 26%. Most respondents held a Bachelor’s degree (76%). A total of 94% of respondents reported that they had heard of or received information about AVs before. Nevertheless, 53% had never used Advanced Driver Assistance Systems (ADAS). Of respondents, 20% drove every day of the week, while 16% did not drive at all. A total of 74% had no problems driving a car, and 72% owned two or more cars, as shown in

Table 2. Demographic and driving information summary.

Variable	Category	Frequency (n = 299)	Percentage (%)
Gender	Male	165	55
	Female	134	45
Age (Years)	18–20	25	8
	21–29	92	31
	30–39	79	26
	40–49	35	12
	≥50	68	23
Level of Education	Lower than	32	11
	Bachelor’s degree	229	76
	Higher than Bachelor’s degree	38	13
Monthly Income	Less than 15,000 THB	74	25
	15,000–25,000 THB	57	19
	25,001–50,000 THB	57	19
	50,001–100,000 THB	87	29
	Higher than 100,000 THB	24	8
Heard of/received information regarding AVs	Yes	280	94
	No	19	6
Using ADAS (Multiple answers)	None	159	53
	Cruise Control	90	30
	Adaptive Cruise Control	73	24
	Lane Keeping Assistance	81	27
	Automated Parking System	62	21
	Blind Spot Detection	69	23
	Forward Collision Warning System	64	21
	Automatic Emergency Braking Systems	41	14
Driving per week	None	47	16
	1–3 days per week	63	21
	4–6 days per week	128	43
	Every day	61	20
Driving problems	Yes	79	26
	No	220	74
Number of cars owned by family	0	13	4
	1	70	24
	≥2	216	72

Note: 1 USD ≈ 32.60 THB at the time of the data collection.

.

3.3. Measurement Development

This research focused on six constructs based on the extended TAM model, including Intention to Use (INT), Perceived Ease of Use (PEOU), Perceived Usefulness (PU), Trust, Perceived Risk (PR), and Environmental Concern (EC). As indicated in

Table 3. Construct measurements.

Constructs	Items	Mean	S.D.	Sources
Perceived Ease of Use (Mean = 3.875, S.D. = 0.833)	1. SAE Level 3 are simple to operate.	4.04	0.956	[12,24]
	2. Operating SAE Level 3 do not require a lot of effort.	3.66	1.015
	3. I can quickly learn how to drive SAE Level 3.	3.92	0.945
Perceived Usefulness (Mean = 3.84, S.D. = 0.807)	4. Using SAE Level 3 will improve my driving efficiency.	3.79	1.072	[12,16,24]
	5. Using SAE Level 3 will improve travel safety.	3.80	1.002
	6. Using SAE Level 3 will reduce the chance of accidents (e.g., decreasing human error.)	3.76	0.994
	7. Using SAE Level 3 will reduce the fatigue from long trips.	4.01	0.980
Intention to Use (Mean = 3.947, S.D. = 0.835)	8. I’d like to try using SAE Level 3 in the future.	4.05	0.936	[26]
	9. If I have the opportunity, I would like to use SAE Level 3 right now.	3.85	0.966
	10. If the pricing is reasonable, I intend to purchase SAE Level 3.	3.94	0.962
Trust (Mean = 3.656, S.D. = 0.862)	11. I trust SAE Level 3 for traveling.	3.59	0.967	[13,27]
	12. I feel comfortable if people who are important to me (e.g., family, close friends) travel by SAE Level 3.	3.67	0.980
	13. I think SAE Level 3 are dependable and functional.	3.70	0.991
	14. Overall, I trust SAE Level 3.	3.67	0.977
Perceived Risk (Mean = 3.744, S.D. = 0.854)	15. Using SAE Level 3will cost me more than usual.	3.69	1.042	[13,27]
	16. SAE Level 3 may not perform well and may experience technical difficulties.	3.79	0.973
	17. Using SAE Level 3 increase the risk of hacking systems and personal data.	3.72	0.986
	18. Using SAE Level 3 makes me concerned about legal liability (e.g., in the event of an accident, who will be responsible).	3.78	1.049
Environmental Concern (Mean = 3.950, S.D. = 0.911)	19. Using SAE Level 3 will help to reduce air pollution.	3.81	1.107	[21]
	20. It is my responsibility to adopt a low-carbon (low-emissions) mode of transportation.	3.91	0.996
	21. Automobile exhaust emission is a major cause of air pollution.	4.13	1.026

Note: the surveys used in this study were provided in the Thai language.

, question items had been developed based on previous research. All questionnaire items were based on a five-point Likert scale ranging from “strongly disagree (=1) to strongly agree (=5)”.

3.4. Data Analysis

The data obtained in this study were analyzed using Statistical Package for the Social Sciences (SPSS) version 23.0 software along with the AMOS version 23.0 software. The methods of exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and structural equation model (SEM) were utilized to investigate the relationships between various psychological variables.

The analyses were conducted in two stages: first, the model’s reliability and validity were checked and then the model hypothesis was tested.

4. Results

4.1. Data Analysis of the Measurement Model

EFA was conducted using component rotation with a fixed number of factors, as in previous literature. Two PU questions (PU4, PU7) were excluded as the recommended value for item loading must be greater than 0.6, as the factor must extract sufficient variance from the variable [28], the Kaiser–Mayer–Olkin (KMO) sampling adequacy score is 0.943, whereas the Bartlett’s test of sphericity is (X²(171) = 3603; 253, p-value < 0.001) indicating that the dataset is adequately sampled and that factor analysis of the data is appropriate, with a recommended communality greater than 0.5 for predicting the variable’s value. The total extracted variance of constructs was 76.8%, as illustrated in

Table 4. Communalities and rotated components.

Factor	Construct	Communalities	Component
			1	2	3	4	5	6
Trust	Trust_Q22	0.773	0.774
	Trust_Q23	0.796	0.769
	Trust_Q20	0.764	0.767
	Trust_Q21	0.780	0.748
PR	PR_Q25	0.775		0.797
	PR_Q27	0.744		0.789
	PR_Q26	0.720		0.756
	PR_Q24	0.666		0.716
PEOU	PEOU_Q2	0.766			0.812
	PEOU_Q1	0.756			0.729
	PEOU_Q3	0.723			0.716
INT	INT_Q18	0.765				0.757
	INT_Q19	0.784				0.725
	INT_Q17	0.796				0.711
EC	EC_Q30	0.809					0.737
	EC_Q28	0.787					0.723
	EC_Q29	0.735					0.653
PU	PU_Q6	0.865						0.775
	PU_Q5	0.792						0.631

Note: PR = Perceived Risk; PEOU = Perceived Ease of Use; INT = Intention; EC = Environmental Concern; PU = Perceived Usefulness.

, confirming that the EFA test was acceptable [29].

The measurement models, including Cronbach’s alpha, were checked to validate internal consistency, which was greater than 0.7, and then the convergent and discriminant validities were checked. The composite reliability (CR) must be greater than 0.7 to validate convergent validity [30,31], and the average variance extracted (AVE) must be greater than 0.5 to validate constructs [32,33]. Finally, the heterotrait–monotrait ratio of correlations (HTMT) technique was utilized to evaluate the discriminant validity of the measurement structure with recommended value less than 0.9 [34]. Cronbach’s alpha, item loading, CR, and AVE values are reported in

Table 5. Reliability and validity assessments.

Factor	Construct	Factoring Loading	Cronbach’s Alpha	CR	AVE
Trust	Trust_Q22	0.859	0.903	0.904	0.701
	Trust_Q23	0.823
	Trust_Q20	0.858
	Trust_Q21	0.808
PR	PR_Q25	0.807	0.864	0.866	0.618
	PR_Q27	0.783
	PR_Q26	0.817
	PR_Q24	0.734
PEOU	PEOU_Q2	0.779	0.820	0.820	0.604
	PEOU_Q1	0.724
	PEOU_Q3	0.826
INT	INT_Q18	0.822	0.845	0.846	0.648
	INT_Q19	0.746
	INT_Q17	0.844
EC	EC_Q30	0.815	0.844	0.844	0.644
	EC_Q28	0.819
	EC_Q29	0.772
PU	PU_Q6	0.743	0.752	0.753	0.605
	PU_Q5	0.811

Note: CR = Composite Reliability; AVE = Average Variance Extracted.

, and discriminant validity is shown in

Table 6. Discriminant validity.

Factor	Trust	PR	PEOU	INT	EC	PU
Trust
PR	0.587
PEOU	0.6795	0.5612
INT	0.7647	0.6353	0.714
EC	0.6614	0.7973	0.6947	0.7075
PU	0.817	0.5543	0.8084	0.6894	0.7019

.

4.2. Data Analysis of the Structural Model

The results of the structural equation model analysis using the AMOS program were used to test the study’s hypothesis. The Chi-square (χ²) was 252.165; the cmin/df was 1.801, indicating good fit; the root mean square error of approximation (RMSEA) was 0.052, representing that our model fits a population; the comparative fit index (CFI) was 0.968, indicating better fit with high values; the root mean residual (RMR) was 0.053; and the standardized root mean squared residual (SRMR) was 0.0528, representing better fit with low values, as given in

Table 7. Model fit.

Indices	χ2	df	p-Value	Cmin/df	RMSEA	TLI	CFI	RMR	SRMR
Good fit	-	-	>0.05	≤2	≤0.05	≥0.95	≥0.95	≤0.05	≤0.05
Acceptable fit	-	-	-	≤3	≤0.08	-	-	>0.05	>0.05
Model	252.165	140	0.000	1.801	0.052	0.961	0.968	0.053	0.0528

, confirming that the model was acceptable.

The proposed model explained 69% of the variance in intention, 79% of the variance in PEOU, and 50% of the variance in PU (see Figure 2). The analytical results based on the hypothesis proposed are as shown in

Table 8. Hypotheses tests.

Hypothesis	Path	β	t Value	Significance	Yes or No
H1	PU→INT	−0.148	−0.929	0.353	No
H2	PEOU→INT	0.306	2.941	0.003	Yes
H3	PEOU→PU	0.406	4.901	***	Yes
H4	Trust→INT	0.482	4.051	***	Yes
H5	Trust→PU	0.460	4.678	***	Yes
H6	Trust→PEOU	0.704	10.578	***	Yes
H7	PR→INT	0.125	1.362	0.173	No
H8	EC→INT	0.192	1.780	0.075	No
H9	EC→PU	0.127	1.740	0.082	No

Note: *** p < 0.001.

and are summarized as follow:

H1. 

‘Perceived usefulness positively affects the intention to use SAE Level 3’ was rejected at the 0.05 level of significant (β = −0.148, p = 0.353).

H2. 

‘Perceived ease of use positively affects the intention to use SAE Level 3’ was accepted at the 0.05 level of significant (β = 0.306, p = 0.003).

H3. 

‘Perceived ease of use positively affects the perceived usefulness of SAE Level 3’ was accepted at the 0.05 level of significant (β = 0.406, p < 0.001).

H4. 

‘Trust positively affects the intention to use SAE Level 3’ was accepted at the 0.05 level of significant (β = 0.482, p < 0.001).

H5. 

‘Trust positively affects the perceived usefulness of SAE Level 3’ was accepted at the 0.05 level of significant (β = 0.460, p < 0.001).

H6. 

‘Trust positively affects the perceived ease of use of SAE Level 3’ was accepted at the 0.05 level of significant (β = 0.704, p < 0.001).

H7. 

‘Perceived risk negatively affects the intention to use SAE Level 3’ was rejected at the 0.05 level of significant (β = 0.125, p = 0.173).

H8. 

‘Environmental concern positively affects the intention to use SAE Level 3’ was rejected at the 0.05 level of significant (β = 0.192, p = 0.075).

H9. 

‘Environmental concern positively affects the perceived usefulness of SAE Level 3’ was rejected at the 0.05 level of significant (β = 0.127, p = 0.082).

5. Discussion

The investigations of TAM and other factors; including Trust, PR, and EC, on the adoption of SAE Level 3 were explored in this research. Based on the findings, the following factors were explained and discussed:

First, unlike previous studies, it was found in this study that the PU factor had no effect on intention to use SAE Level 3. The descriptive analysis in this study showed that 94% of respondents had heard about AVs before and 72% of them had two or more cars in their households. Nevertheless, less than half of the respondents had utilized the ADAS system. This indicates that some of the respondents had not experienced ADAS or that their vehicles were not equipped with the ADAS technology. Therefore, these individuals are less likely to understand the benefits of SAE Level 3. Furthermore, as reported by T. Thananusak et al. (2017), the primary concern of Thai car consumers was speed [35]. Therefore, it can be seen that Thai people buy a car mainly on its overall performance rather than the benefits of a specific technology.

Second, it was found that the PEOU influenced the intention to use SAE Level 3. This is consistent with previous research regarding the adoption of SAE Level 3 [15,36]. Panagiotopoulos and Dimitrakopoulos (2018) conducted research on SAE Level 5 and suggested that although the PEOU factor had little effect on intention, AVs developers needed to focus more on the system’s friendliness [20]. Additionally, Xu et al. (2018) revealed that PEOU significantly predicted behavioral intention only when participants were riding in SAE Level 3 [15]. It could be argued that a SAE Level 3 still required human drivers, and therefore less effort in learning how the technology works was necessary. Thus, in terms of promoting AVs in the future, car manufacturers or dealers should provide more demonstrations to users for a better understanding of how to use AVs.

Moreover, our findings revealed that the PEOU factor had a significant effect on PU, which is in line with previous research regarding the adoption of SAE Level 3–5 [16,20,23,36]. Hence, to encourage the upcoming adoption of automated cars, automakers should design AVs systems in such a way that more individuals see the ease of AV use and the benefits of AVs.

Third, the intention to use new or unfamiliar technology can be influenced by the Trust factor [27,37]. Our study indicated that the strongest factor affecting the intention to use SAE Level 3 was found to be Trust. This is in line with previous studies that used questionnaire survey, which discovered that Trust was a crucial variable determining SAE Level 3 usage intention [13,36]. This is also supported by the results of driving simulator tests, which showed that Trust was a major determinant of willingness to engage in SAE Level 3 [38]. In addition, I. Nastjuk et al. (2020) discovered that operating SAE Level 5 could help users have a better driving experience if they understood how autonomous driving systems work [22]. Another research showed that perceptions of the system’s reliability influenced the decision to adopt SAE Level 5 [26]. As reported by ABeam Thailand (ABeam) from a study on car buying behavior of consumers in the Thai market in 2019 and 2021, the quality of products, services, and customer relationships all played a big part in brand loyalty, with 93% of car buyers in Thailand not changing their decision about buying the brand they preferred during the COVID-19 pandemic [39]. Although AVs have undergone extensively thorough testing for the safety of driving systems, even the slightest error can damage users’ trust. Therefore, to promote the use of automation driving technology, car manufacturers and dealers should focus on the car’s reliability and after-sales services, as well as customer relationships.

As for the impact of Trust on PU and PEOU, we discovered that Trust has a direct effect on the PU and PEOU constructs, which conforms to the findings of Xu et al. (2018), who discovered that the Trust factor has a direct influence on PU and PEOU [15]. Therefore, the greater trust in SAE Level 3, the greater the Perceptions of Usefulness and Ease of Use. The government or dealers and manufacturers should thus promote, for example, giving individuals the opportunity to use AVs to gain more experience and familiarity with the technology, educating Thai people about the benefits of AVs (e.g., reduction of fuel consumption, severe injuries, fatigue), and improving public facilities, such as charging stations, etc.

Forth, regarding the Perceived Risk factor, this study revealed that PR had no impact on the peoples’ intentions to use SAE Level 3. This might be due to the respondents’ lack of knowledge regarding the cost of SAE Level 3, which may be more expensive than a conventional car or may have higher maintenance costs. Furthermore, as they have not driven SAE Level 3 before, they may be unaware that the car may have usability concerns and technical difficulties. As for hacking systems, it could be possible that the respondents were not concerned about the security of SAE Level 3. This could be because SAE Level 3 data processing and communications are not as automatic as SAE Level 5, which raise far more concerns about cybersecurity and data privacy.

Lastly, as for the EC construct, our investigation indicated that EC had no significant influence on the intention to use SAE Level 3 and PU. However, according to the research by Wu et al. (2019) [21], environmentally conscious people intend to use electric AVs (implied to SAE Level 4–5). Moreover, this study only considers factors that are relevant to environmental concerns and mainly focuses on the investigation of AVs that are powered by electricity. However, a number of factors affected intention, and some SAE Level 3 vehicles might be powered by sources other than electricity, such hydrogen or fossil fuels. Furthermore, since SAE Level 3 have not been widely used in Thailand, it might be difficult for the public to understand its environmental benefits. Consequently, it could be stated that when it comes to purchasing a SAE Level 3, Thai buyers prioritize other considerations over environmental concerns.

6. Conclusions and Recommendations

This study investigated the factors that affect individuals’ intention to use SAE Level 3 by bringing in TAM theory and integrating factors such as Trust, PR, and EC into the theoretical model. The main objective was to examine which factors influenced people’s intentions to use SAE Level 3 with the high hope that the findings could help encourage the use of AVs in the future.

This study showed that the intention to use SAE Level 3 was strongly influenced by Trust, followed by Perceived Ease of Use (PEOU). In addition, it was found that Trust affected PEOU and Perceived Usefulness (PU). Although PU was affected by PEOU, it had no effect on adoption intention. These findings implied that Thai car buyers choose cars based on their performance rather than their benefits.

Our findings suggest that future policies aimed at encouraging the adoption of AVs might involve informational campaigns emphasizing the benefits of AVs to build public trust. Additionally, infrastructure upgrades should be targeted to encourage more individuals to utilize AVs (e.g., charging stations, roads, and traffic systems). Similarly, technology developers and car dealers should prioritize the user-friendliness of AV systems, AV demonstrations to increase reliability, and after-sales services to foster customer trust.

The results of this study have to be interpreted while bearing in mind certain limitations. First, the participants were not perfectly controlled; about 16% of the participants do not drive weekly. However, they could be potential adopters of the AVs market. Second, there may be a concern that the distribution of the questionnaire via Facebook could bias the answers in favor of individuals who use technology. However, in Thailand, Facebook has been widely used for over ten years, so this channel of survey form distribution is able to reach every group of people.

Future research may investigate specific groups of prospective first-time car buyers (e.g., young people (who have kept up with trends and technology)) and car owners who may purchase more cars (e.g., working age adults or the elderly (who can afford AVs)), and assess the possibility of adopting shared AVs or autonomous public transportation as a policy to facilitate elderly people who use public transit because it is safer than driving on their own [40]. Future studies may periodically be conducted comparing the changes over the next 5 to 10 years to the adoption of AVs. Moreover, the factors that influence Trust should be explored, for instance, system transparency, technical competence, situation management, subjective norms, and self-efficacy [13,26]. Alternatively, the inclusion of the social media factor should be considered, as a study conducted in Thailand showed that social media can be effectively utilized for travel planning [41]. In the future, therefore, this could be used for trip planning with AVs.