Journal of Entrepreneurship, Management and Innovation (2025)

Volume 21 Issue 4: 30-53

DOI: https://doi.org/10.7341/20252142

JEL Codes: L26, M13, O31

Son Tung Ha, Ph.D., Associate Professor at the Faculty of Business Management, School of Business, National Economics University, 207 Giai Phong, Hanoi, Vietnam, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Thi Thanh Hoa Phan, Ph.D., Faculty of Business Management, School of Business, National Economics University, 207 Giai Phong, Hanoi, Vietnam, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Thi Viet Nga Ngo, Ph.D. at the Faculty of Business Management, School of Business, National Economics University, 207 Giai Phong, Hanoi, Vietnam. Email: This email address is being protected from spambots. You need JavaScript enabled to view it.
Cong Doanh Duong, Ph.D., Associate Professor at the Faculty of Business Management, School of Business, National Economics University, 207 Giai Phong, Hanoi, Vietnam,
e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Ngoc Thang Ha, Ph.D., Faculty of Business Management, School of Business, National Economics University, 207 Giai Phong, Hanoi, Vietnam. Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

PURPOSE: Sustainability-oriented entrepreneurship plays a pivotal role in addressing global environmental and social challenges by aligning economic activity with sustainable development goals. While the theory of planned behavior has been widely applied to explain entrepreneurial intentions, limited attention has been given to the influence of artificial intelligence-related competencies on such intentions. This study aims to examine how knowledge of artificial intelligence and confidence in using artificial intelligence tools influence university students’ intentions to engage in sustainability-oriented entrepreneurship, thereby extending the theory of planned behavior. METHODOLOGY: A cross-sectional survey was conducted with a sample of 217 undergraduate students from five universities in Vietnam, selected using a stratified random sampling approach. Multiple linear regression was used to test the direct effect, while the PROCESS macro approach was employed to test the mediation effect. Polynomial regression and response surface analysis were employed to investigate how attitudes towards sustainability-oriented entrepreneurship and perceived behavioral control are congruent or incongruent with each other in triggering sustainability-oriented entrepreneurial intentions. FINDINGS: The results demonstrate that a positive attitude toward sustainability-oriented entrepreneurship and a strong sense of control over entrepreneurial actions are significant predictors of entrepreneurial intentions. Intentions are highest when both attitude and perceived behavioral control are simultaneously strong, indicating a synergistic effect. However, imbalances between these two factors do not significantly reduce intention. Knowledge of artificial intelligence and self-confidence in using AI tools. Moreover, subjective norms do not directly influence intentions. IMPLICATIONS: The study advances theoretical understanding by incorporating emerging technological competencies into the theory of planned behavior framework. For practitioners and educators, the findings suggest that enhancing artificial intelligence capabilities among students may indirectly foster stronger intentions to engage in sustainability-oriented entrepreneurship. ORIGINALITY AND VALUE: This research is among the first to integrate artificial intelligence-related constructs into a well-established psychological framework for explaining sustainable entrepreneurship. It offers novel insights into how technological competencies contribute to entrepreneurial motivation through established cognitive pathways.

Keywords: AI knowledge, AI self-efficacy, sustainability-oriented entrepreneurial intention, theory of planned behavior, response surface analysis, artificial intelligence competencies, sustainability-oriented entrepreneurship, entrepreneurial intentions, sustainable development, university students, technological competencies, attitude–behavior–control framework, mediation analysis

INTRODUCTION

Sustainability-oriented entrepreneurship (SOE) has become a crucial strategy for confronting pressing worldwide issues such as climate change, resource scarcity, and social inequity (Srivastava et al., 2024a). It differs from traditional entrepreneurship by emphasizing a triple-bottom-line approach that integrates economic, environmental, and social objectives (Hopp et al., 2025). SOE is broadly defined as pursuing entrepreneurial ventures to achieve sustainable development while balancing profitability with societal and environmental welfare (Lazarte-Aguirre, 2024). In recent years, this form of entrepreneurship has gained increasing prominence as a means to contribute to the United Nations’ Sustainable Development Goals (SDGs) through innovative solutions to socio-environmental problems (Agu et al., 2021). This trend reflects the critical role of SOE in driving long-term societal change and highlights the need to understand what motivates individuals to engage in such ventures. Although the importance of SOE is widely recognized globally, there remains a knowledge gap regarding the factors that encourage aspiring entrepreneurs to adopt sustainability-oriented initiatives (Srivastava et al., 2024a). Bridging this gap is particularly relevant in emerging economies like Vietnam, where rapid industrialization has led to complex sustainability challenges and a pressing need for entrepreneurial solutions (Duong, 2025). Understanding the drivers of SOE intentions in such contexts is crucial for advancing both theory and practice in sustainable entrepreneurship.

Artificial intelligence (AI) has rapidly become a transformative force in entrepreneurship, offering new avenues to support and enhance SOE (Al-Romeedy & El-Sisi, 2024). Entrepreneurs increasingly leverage AI technologies – including machine learning, predictive analytics, and generative AI – to process large volumes of data, gain actionable insights, and improve decision-making (Bickley et al., 2024). These capabilities can significantly benefit sustainability-focused ventures. AI-driven analytics enable entrepreneurs to identify sustainable opportunities and innovative business models that might remain hidden from human analysis alone. By reducing uncertainty in highly dynamic markets, AI tools help entrepreneurs better evaluate the feasibility of projects that balance economic and environmental goals (Duong & Nguyen, 2024). In practical terms, AI applications have been shown to optimize energy usage, improve waste management, and promote circular economy practices – all of which directly contribute to the environmental mission of SOE (Gupta et al., 2023). Moreover, AI can streamline operations and automate routine tasks, freeing entrepreneurs to focus on the creative and strategic aspects of their sustainable venture (Roundy, 2022). By providing data-driven evidence of success, AI may strengthen an individual’s positive attitude toward starting a sustainable enterprise and their perceived control over achieving sustainability goals (Bickley et al., 2024). Thus, emerging research suggests that AI technology contributes directly to sustainable outcomes and influences entrepreneurs’ mindsets, making sustainable entrepreneurship more appealing and attainable (Bonfanti et al., 2024).

However, this optimistic view of AI warrants greater scrutiny. Critics point out that AI technologies themselves consume considerable energy and resources, raising concerns about their environmental footprint (Ueda et al., 2024). Moreover, access to AI tools is often unequal, exacerbating digital divides and creating barriers for under-resourced entrepreneurs (Imjaj et al., 2025). These ethical and structural challenges highlight the need to consider both enabling and constraining roles of AI in sustainable entrepreneurship. To date, few studies have critically examined these dualities, resulting in a gap in understanding the full implications of AI for sustainability. Our study acknowledges these tensions and focuses on the cognitive influence of AI-related competencies while recognizing the broader socio-technical discourse. Moreover, despite growing interest in digital entrepreneurship and sustainability (Duong, 2025), little attention is paid to the role of such technological enablers in shaping sustainability-oriented entrepreneurial intentions.

In considering the influence of AI on SOE at the individual level, two personal factors are particularly salient: AI knowledge and AI self-efficacy. AI knowledge refers to an individual’s understanding of and proficiency with AI tools (Chiu et al., 2024) and concepts relevant to business (Imjai et al., 2024), while AI self-efficacy denotes one’s confidence in their ability to use AI technologies effectively (Bewersdorff et al., 2025). AI-literate and AI-confident individuals could have an edge in recognizing and acting on sustainable opportunities, suggesting a direct positive effect of these qualities on their entrepreneurial intentions. These expectations align with broader findings in entrepreneurship research that domain-specific knowledge and self-efficacy bolster entrepreneurial intent and action (Al Issa et al., 2025; Renko et al., 2012). Nonetheless, the extent to which AI knowledge and self-efficacy independently drive SOE intentions remains to be empirically tested. To address the aforementioned gaps, our study adopts the theory of planned behavior (TPB) (Ajzen, 1991) to explore how AI drivers (AI knowledge and AI self-efficacy) foster individuals’ sustainability-oriented entrepreneurial intention underlying cognitive mechanisms formed by core components in the TPB, including attitude towards sustainability-oriented entrepreneurship (ATS) and perceived behavioral control (PBC), and subjective norms (SN). Particularly, this study aims to address the following research questions (RQs):

RQ1: How do AI knowledge and AI self-efficacy influence the cognitive antecedents of sustainability-oriented

entrepreneurial intentions (attitude toward SOE and perceived behavioral control) and, ultimately, sustainability-oriented entrepreneurial intentions themselves?

RQ2: How do attitudes toward SOE and perceived behavioral control interact under congruent and incongruent conditions

(i.e., when both are similarly high or low versus when one is high and the other low) to affect sustainability-oriented entrepreneurial intentions?

RQ3: To what extent do attitude and perceived behavioral control mediate the relationship between AI-related factors (AI

knowledge and AI self-efficacy) and sustainability-oriented entrepreneurial intentions?

The remainder of this paper reviews the relevant literature and develops the research hypotheses. This is followed by a description of the methodology, including data collection procedures, measurement instruments, and analytical techniques. The subsequent section presents the empirical results, which are then discussed in relation to existing literature. Finally, the paper concludes by outlining key theoretical and practical implications, acknowledging limitations, and offering directions for future research.

LITERATURE REVIEW

Sustainability-oriented entrepreneurship

SOE refers to entrepreneurial activities that explicitly aim to achieve sustainable development goals by balancing economic success with positive environmental and social impact (Duong, 2025). SOE extends the traditional profit-oriented venture creation process to encompass a broader set of objectives, often described as the triple bottom line of “people, planet, and profit” (Bonfanti et al., 2024). Entrepreneurs in this domain seek innovative solutions to problems such as climate change, resource scarcity, and social inequality, integrating sustainability considerations into their business models from inception. This approach to entrepreneurship has gained momentum as stakeholders increasingly call for businesses to contribute to sustainable development (Srivastava et al., 2024a). Governments and international bodies encourage entrepreneurial action on sustainability issues, recognizing SOE as a means to help meet targets like the United Nations Sustainable Development Goals (Kwilinski et al., 2024). Likewise, consumers have shown rising interest in supporting companies with environmental and social missions, creating market opportunities for sustainability-driven startups (Bellver et al., 2022). By addressing urgent global challenges through enterprise, SOE is crucial in facilitating systemic change toward sustainability and creating economic value (Ip, 2024). This dual value creation is what sets SOE apart from conventional entrepreneurship, which has historically prioritized financial performance, often at the expense of environmental or social considerations.

Research on sustainability-oriented entrepreneurship has grown over the past decade, yielding insights into the motivations and factors that drive individuals to engage in this form of venture creation. One line of inquiry has examined sustainable entrepreneurs’ personal values and ethics. Studies suggest that entrepreneurs who prioritize altruistic, pro-environmental values are more likely to pursue sustainability-oriented ventures, even when uncertain profit potential (Kuckertz & Wagner, 2010). For instance, Vuorio et al. (2017) found that among Finnish entrepreneurs, biospheric values (concern for nature and the ecosystem) significantly influenced sustainability-oriented entrepreneurial intentions, primarily via shaping positive attitudes toward sustainable business. Another line of research has focused on cognitive and educational factors. Entrepreneurial intention models, including the TPB, have been applied in the context of sustainability to identify key antecedents of intentions to start a sustainable enterprise (Romero-Colmenares & Reyes-Rodríguez, 2022). These studies confirm that the same psychological drivers are important in general entrepreneurship and relevant for SOE, though sometimes with differing magnitudes.

Some work shows that perceived social pressure (subjective norms) can be particularly salient for social or sustainable ventures due to normative expectations around “doing good,” whereas in classic profit-driven entrepreneurship, it might be less critical (Sharma, Bulsara, Bagdi, et al., 2023; Srivastava et al., 2024b). Other research emphasizes self-perception of capability: individuals are more inclined to start a sustainable venture if they feel competent in entrepreneurship and knowledgeable about sustainability issues (Wang et al., 2023). An important contextual factor is entrepreneurship education with a sustainability focus. Exposure to sustainability concepts in academic or training programs has been shown to raise awareness of environmental problems and enhance students’ commitment to addressing them through entrepreneurship (Alimehmeti et al., 2025). In a recent study of university students, those who received education in sustainable development and social responsibility reported higher sustainable entrepreneurial intentions, mediated by stronger pro-sustainability attitudes (Duong, 2025). Moreover, there is ongoing debate regarding the limitations of technology-driven approaches to sustainability, particularly concerning unintended consequences, rebound effects, or the risk that digital and AI tools are deployed in ways that ultimately support unsustainable business models.

Recent years have witnessed a surge of interest in how digital technologies—and particularly artificial intelligence—can drive or support sustainable entrepreneurship. Studies have shown that digital skills, information and communication technology (ICT) applications, and data analytics capabilities are associated with greater entrepreneurial intention and success (Duong et al., 2024; Fazio et al., 2024). For instance, mastery of ICT tools can help entrepreneurs identify market opportunities, design innovative products, and reach new customer segments, while data-driven decision-making can enhance operational efficiency and strategic agility. In the context of sustainability, AI can facilitate resource optimization, enable the tracking and reduction of environmental impact, and support the development of new business models such as the circular economy (Giuggioli & Pellegrini, 2022; Imjai et al., 2024). However, there is also evidence that technological sophistication alone is insufficient to guarantee positive sustainability outcomes.

Critics have raised important concerns regarding the ethical, environmental, and social risks of digital technologies. For example, AI systems can consume significant amounts of energy, contribute to electronic waste, or be harnessed for purposes that are at odds with social or environmental goals (Guo et al., 2025). The diffusion of AI and other advanced technologies may also exacerbate existing inequalities or lead to unintended negative consequences if not aligned with broader sustainability principles (Vinuesa et al., 2020). These counter-narratives underscore the importance of adopting a critical and holistic perspective when assessing the role of AI in SOE, acknowledging both its enabling and potentially problematic aspects. Empirically, research remains limited on the direct and indirect pathways through which AI knowledge and self-efficacy influence sustainability-oriented entrepreneurial intention, particularly in understanding how these effects manifest within different institutional, social, and cultural contexts. This study aims to address these gaps by not only examining individual cognitive mechanisms but also acknowledging the broader social and technological landscape in which SOE emerges.

Theory of planned behavior

The TPB provides a robust theoretical lens for understanding entrepreneurial intentions (Ahmed et al., 2025), including those oriented toward sustainability (Sharma, Bulsara, Bagdi, et al., 2023). According to Ajzen (1991), the intention to perform a behavior is jointly determined by: Attitude toward the behavior – the individual’s overall evaluation (favorable or unfavorable) of performing the behavior; Subjective norms – the perceived social pressures or support to perform (or not perform) the behavior; and Perceived behavioral control– the perceived ease or difficulty of performing the behavior, which reflects one’s sense of capability and control over the action. TPB posits that a more positive attitude, stronger supportive norms, and higher perceived control will each contribute to a stronger intention to engage in the behavior (Relente & Capistrano, 2024). The applicability of TPB to entrepreneurship is well-established: entrepreneurship is considered a planned, intentional act that often involves significant forethought and personal agency (Krueger et al., 2000). Indeed, TPB has been one of the most widely used models to predict entrepreneurial intentions (Ahmed et al., 2025; Relente & Capistrano, 2024). Meta-analyses and reviews show that the TPB constructs collectively explain a substantial portion of the variance in intentions to start a business (Zaremohzzabieh et al., 2019), making it a useful framework for identifying key motivational levers.

Despite its value, the TPB has also attracted criticism for its focus on individual-level cognitive antecedents, with less emphasis on external or contextual factors such as institutional support, policy, or collective action (Abu Shriha et al., 2024; Ahmed et al., 2025). Some scholars argue that entrepreneurial intentions—and especially those relating to sustainability—are not only shaped by attitudes, subjective norms, and perceived control, but also by the broader structures and systems within which entrepreneurs operate (Lortie & Castogiovanni, 2015). As such, integrating perspectives from institutional theory, practice theory, or socio-technical transitions literature could enrich understanding of how individual intentions interact with enabling or constraining contexts. In the specific context of sustainable entrepreneurship, TPB has also proven insightful. Sharma, Bulsara, Bagdi, et al. (2023) applied TPB to study sustainable entrepreneurial intentions and found that attitudes towards sustainable entrepreneurship and PBC were significant predictors of students’ intentions to create sustainable enterprises, while subjective norms had a smaller effect. Similar results were reported by Heredia-Carroza et al. (2024) in a study on entrepreneurship intentions, where personal attitudes and perceived behavioral control were found to be crucial, while social norms were not. These findings suggest that while all three TPB factors can matter, individuals’ own positive evaluations and confidence in performing sustainable entrepreneurship may be especially critical in driving their intention.

A person with a favorable attitude towards starting a sustainable venture is likely to view such entrepreneurship as personally rewarding or worthwhile, which should increase their intention to pursue it. By contrast, if they view sustainable entrepreneurship as unattractive or unimportant, their intention will be weaker. Prior research consistently shows that attitude is one of the strongest predictors of entrepreneurial intention in general (Ahmed et al., 2025; Relente & Capistrano, 2024). In sustainable entrepreneurship, attitude often encapsulates one’s commitment to sustainability values and excitement about entrepreneurial opportunity. When individuals genuinely care about sustainable development and believe that creating a venture is a meaningful way to contribute, they are more inclined to form intentions to do so (Waris et al., 2021). For example, a student who is passionate about clean energy and sees business as an effective vehicle for impact will likely hold a strong positive attitude towards launching a solar energy startup, which drives their intention. Empirical studies support this logic: attitudes have been found to significantly and positively correlate with sustainability-oriented entrepreneurial intentions (Sharma, Bulsara, Trivedi, et al., 2023; Vuorio et al., 2017). Therefore, we expect a similar positive relationship in our context.

H1: Attitude towards sustainability-oriented entrepreneurship is positively correlated with sustainability-oriented

entrepreneurial intentions.

According to TPB, greater PBC should strengthen intention, especially for behaviors that require skill and effort (Ajzen, 1991). In entrepreneurial settings, PBC is often operationalized similarly to entrepreneurial self-efficacy – belief in one’s entrepreneurial capabilities (Zhao et al., 2005). Numerous studies have demonstrated that entrepreneurial self-efficacy, or PBC, strongly predicts the intention to start a business (Krueger et al., 2000; Neneh, 2020). The reason is intuitive: those who feel competent and in control are more willing to initiate entrepreneurial action, as they expect to manage the process effectively, whereas those who doubt their abilities are hesitant to commit. In the sustainability context, perceived behavioral control may encompass general business skills and the perceived ability to achieve sustainability outcomes (Sharma, Bulsara, Trivedi, et al., 2023; Vuorio et al., 2017). Tan et al. (2020) found that perceived feasibility (akin to PBC) of social entrepreneurship significantly contributed to intention among participants – when individuals believed they had the know-how and resources to start a social enterprise, their intentions solidified. Likewise, in green entrepreneurship research, PBC (such as confidence in implementing green practices or obtaining needed resources) correlates with stronger entrepreneurial intentions (Srivastava et al., 2024a). Thus, we anticipate that students who feel a high sense of control over founding a sustainability-oriented venture will be more inclined to do so.

H2: Perceived behavioral control is positively correlated with sustainability-oriented entrepreneurial intentions.

As several scholars have noted, sustainability-oriented entrepreneurship is not just an individual act but is embedded within social contexts, community values, and shared norms (Truong et al., 2022). The omission of subjective norms risks ignoring how perceived support from family, peers, and professional networks may shape the intention to pursue sustainability goals. This is particularly salient in collectivist cultures such as Vietnam, where social conformity and communal expectations often play a central role in decision-making processes (Duong, 2024). In entrepreneurial research, subjective norms may operate as a gatekeeping mechanism by signaling whether a new venture is seen as legitimate or supported by relevant reference groups (Ahmed et al., 2025). In the sustainability domain, such norms might involve approval from environmental peers, educators, or professional mentors who advocate for sustainable innovation (Truong et al., 2022). Thus, including subjective norms enhances the theoretical fidelity of TPB and enables a more nuanced understanding of intention formation. Recent studies have also reported meaningful associations between subjective norms and sustainability-related behavioral intentions. For instance, Truong et al. (2022) found that students’ perception of peer and family support significantly shaped their willingness to engage in social enterprises. Likewise, in the context of green and prosocial entrepreneurship, social expectations have been shown to exert a direct influence on intention, particularly in societies with strong communal values (Srivastava et al., 2024a). When individuals believe that important referents—such as family, friends, or academic mentors—support their pursuit of sustainability-oriented ventures, they are more likely to develop strong entrepreneurial intentions. This may be especially relevant in collectivist societies, where conformity to social expectations and familial encouragement play a central motivational role (Hofstede, 2001). Consequently, we posit that:

H3: Subjective norms are positively correlated with sustainability-oriented entrepreneurial intentions.

Recent theoretical arguments suggest examining how these two factors might combine to influence entrepreneurial intentions (Duong, 2025). Both attitude and PBC are essential, but is having both at strong levels especially potent? Conversely, if one is high and the other low, does that mismatch undermine one’s intention? In entrepreneurial behavior models, there is an implicit notion that for someone to intend and ultimately act, they should both want to do it (attitude/desirability) and feel able to do it (PBC/feasibility) (Krueger et al., 2000; Shapero & Sokol, 1982). When both conditions are met – that is, an individual highly values the sustainable venture and also feels highly capable of executing it – the intention to proceed should be strongest. This situation can be described as a congruence of high attitude and high PBC. On the other hand, if there is an incongruence – for example, the person is very enthusiastic about the idea (high attitude) but has low confidence in their ability, or vice versa – the lack of one component may temper their overall intention. Individuals who love the concept of a sustainable startup but feel unqualified or powerless may hesitate to form a firm intention because they foresee difficulties.

Similarly, someone who is confident in their entrepreneurial skills but not personally invested in sustainability may not be strongly motivated to start a sustainable venture, as the drive or purpose is missing. Therefore, their interplay could matter beyond the additive effects of attitude and PBC. We expect that when both attitude and PBC increase together (congruent increase), they reinforce each other, resulting in especially elevated SOE intentions. In contrast, when one increases without the other (incongruent) – essentially a misalignment – the positive effect of the one might be offset by the deficiency of the other, potentially leading to a reduction in intentions compared to the congruent case. It is important to note that Ajzen (1991)’s TPB model is linear and does not explicitly posit an interaction between attitude and PBC; however, the entrepreneurial context invites the possibility that both favorable mindset and capability perceptions need to align for maximal intention. This perspective aligns with the entrepreneurial event model (Shapero & Sokol, 1982), which argued that both perceived desirability (similar to attitude) and perceived feasibility (similar to PBC) are required triggers for new venture initiation. We, therefore, hypothesize an interaction effect in terms of congruence/incongruence:

H4: A congruent increase in attitude towards sustainability-oriented entrepreneurship and perceived behavioral control

would increase sustainability-oriented entrepreneurial intentions.

H5: An incongruent increase in attitude towards sustainability-oriented entrepreneurship and perceived behavioral

control would reduce sustainability-oriented entrepreneurial intentions.

The role of AI knowledge

While TPB identifies proximal determinants of intention, it also acknowledges that background factors (such as personal traits, knowledge, and situational influences) affect intentions indirectly through those determinants (Ajzen, 1991). In the context of this study, AI knowledge is one such background factor that may influence sustainability-oriented entrepreneurial intentions by shaping attitude and perceived control. We define AI knowledge as an individual’s awareness of and ability to understand AI tools and their potential applications (Chiu et al., 2024). This includes being aware of the available AI technologies, understanding how they can be utilized, and having some experience or proficiency in using them. Individuals vary widely in their AI knowledge: some tech-savvy students might be familiar with machine learning algorithms or have used AI platforms (like predictive analytics or content creation tools), whereas others have only a cursory awareness of AI. We propose that those with more excellent AI knowledge will exhibit more positive attitudes toward engaging in sustainability-oriented entrepreneurship and a higher sense of control, which can spur their entrepreneurial intentions.

Firstly, AI knowledge can enhance attitudes toward SOE by illuminating AI’s possibilities for achieving sustainability goals. A knowledgeable individual is more likely to recognize how AI can make a sustainable venture more effective or innovative (Füller et al., 2022). For example, they might see that using AI for energy management in buildings can significantly cut emissions or that AI-driven analysis of farming data can improve crop yields in an eco-friendly way. This recognition can lead to a more favorable evaluation of starting a venture that leverages AI for sustainability – essentially, they perceive the venture as more promising and impactful. Moreover, familiarity with AI might reduce fear or skepticism about technology in business, leading to a more enthusiastic outlook. Someone who can “speak the language” of AI is apt to have higher confidence in the benefits it can bring, thereby bolstering the appeal (attitude) of an AI-enhanced sustainable business (Gupta et al., 2023). In contrast, individuals with low AI knowledge may feel that integrating advanced tech into a startup is daunting or may underestimate the potential upside, potentially dampening their attitude toward such an endeavor.

Secondly, AI knowledge is likely to increase perceived behavioral control. Knowledge is power: knowing how to use AI tools or at least understanding their logic can make the task of starting a tech-enabled sustainable venture seem more attainable. For instance, a founder who knows how to utilize AI for supply chain optimization will feel more capable of managing operational challenges in an eco-commerce startup. They may anticipate fewer difficulties in implementing AI solutions or be better prepared for troubleshooting issues, contributing to a sense of control. Prior research on technological competencies in entrepreneurship supports this link – entrepreneurs with greater technical knowledge often report higher self-efficacy in related entrepreneurial tasks (Hsu et al., 2019; van der Westhuizen & Goyayi, 2019). In our context, AI knowledge can serve as a resource that lowers perceived barriers: the individual might think, “I know how AI works, so I can handle that aspect of the business”, thus elevating their PBC regarding a sustainability-oriented venture that uses AI.

Beyond these indirect pathways, one might expect that higher AI knowledge could also have a direct positive effect on entrepreneurial intention. If a person is very knowledgeable about AI, they may be inherently more inclined to start a venture exploiting that knowledge – particularly given the current enthusiasm around AI-driven startups. They might see a clear market opportunity or feel a personal drive to capitalize on their AI expertise, which could directly fuel their intention to launch an AI-related sustainable business. While TPB would argue this effect is mediated by attitude and control perceptions, some residual direct influence is possible if, for example, AI knowledge also correlates with other unmodeled factors like an innovative mindset or network access that independently encourage entrepreneurial action. In addition, aligning with the TPB rationale that the influence of background factors on intentions occurs through the proximal factors (Ahmed et al., 2025; Relente & Capistrano, 2024), we expect mediation by attitude and PBC. This mediated relationship acknowledges that even if AI knowledge correlates with intention, the psychological mechanism likely runs through making the individual more optimistic (attitude) and confident (PBC) about starting a sustainable venture. Consequently, drawing these arguments together, we hypothesize a positive relationship between AI knowledge and the key TPB components and outcome:

H6: AI knowledge is positively correlated with (a) attitude towards sustainability-oriented entrepreneurship, (b) perceived

behavioral control, and (c) sustainability-oriented entrepreneurial intentions.

H7: AI knowledge is indirectly correlated with sustainability-oriented entrepreneurial intentions via (a) attitude towards

sustainability-oriented entrepreneurship and (b) perceived behavioral control.

The role of AI self-efficacy

Another crucial AI-related personal factor in our model is AI self-efficacy, which we define as one’s belief in their capability to use and leverage AI technologies for tasks or projects effectively. This concept is rooted in Bandura’s social cognitive theory (Bandura, 1986), where self-efficacy in a specific domain (here, AI) influences how people think, feel, and act in that domain. AI self-efficacy differs from AI knowledge in that it is more about confidence and judgment of one’s abilities than about the factual understanding of AI itself. It is possible to have considerable knowledge yet low confidence (or vice versa), so this construct adds a psychological dimension to the individual’s AI readiness. We posit that AI self-efficacy will shape entrepreneurial attitudes and perceived behavioral control related to SOE, similar to AI knowledge, but through the lens of confidence.

First, high AI self-efficacy should foster a positive attitude toward launching a sustainable venture that uses AI. If individuals feel capable of handling AI tools, they are more likely to favorably view a tech-integrated sustainable business idea. Confidence in using AI can translate into optimism about the outcomes – they trust that they can implement AI solutions successfully, which could make the prospect of the venture more exciting and less intimidating. For instance, a person who is confident in their ability to develop predictive models might be enthusiastic about a sustainable agriculture startup that relies on data analytics to reduce waste because they foresee themselves managing the AI part effectively. This positive outlook feeds into their attitude toward the venture. On the other hand, someone with low AI self-efficacy might harbor doubts, which can impart a more negative or hesitant attitude toward starting that business (Duong, 2025). By analogy, higher AI self-efficacy should engender a more favorable evaluation of an AI-enabled entrepreneurial endeavor.

Second, AI self-efficacy is expected to bolster perceived behavioral control when starting a sustainability-oriented venture strongly. AI self-efficacy can be seen as a domain-specific extension of entrepreneurial self-efficacy. If a person believes “I can use AI tools to solve problems”, this contributes to their overall sense that they can control the tech aspect of their venture. Starting a new business, especially involving advanced technology, often comes with uncertainty about managing all facets. High self-efficacy acts as an assurance: those individuals feel they have what it takes to integrate AI into their business processes, which raises their perceived behavioral control over launching the venture. Prior work shows that self-efficacy in relevant skills tends to increase the perceived feasibility of entrepreneurial activities (Fuller et al., 2018). Latikka et al. (2019) found that people with greater robot-use self-efficacy were more accepting of and open to using AI-driven robots, which implies they felt more in control of interacting with that technology. Translating to our scenario, a budding entrepreneur confident in AI might think, “using AI in my startup is within my abilities”, thereby elevating their PBC. Additionally, AI self-efficacy can reduce anticipated external barriers; confident individuals might be less worried about needing external experts or excessive resources for the AI component, further boosting their sense of control.

Furthermore, prior research indicates that higher levels of self-efficacy enhance an individual’s perception of opportunity recognition and risk management, which are key drivers of entrepreneurial behavior (Aboobaker et al., 2023). Specifically, individuals with strong AI self-efficacy are more likely to view AI not as a source of uncertainty but as a resource that can be strategically leveraged to address sustainability challenges. This perception can strengthen their intention to pursue sustainability-oriented entrepreneurial activities, as they feel more capable of integrating AI solutions into innovative business models. Moreover, AI self-efficacy can mitigate perceived barriers to entrepreneurship by enhancing individuals’ control over technical and operational aspects, which aligns with the determinants of entrepreneurial intentions outlined in the theory of planned behavior (Ajzen, 1991).

Finally, we consider the direct effect: individuals with high AI self-efficacy may be generally more inclined to pursue ventures that involve AI, which could directly nudge their entrepreneurial intentions upward. They might relish the challenge or see it as an opportunity to apply their skills, thereby independently motivating them to plan a startup. Indeed, self-efficacy is known to have a direct positive effect on entrepreneurial intentions in many studies (Al Issa et al., 2025; Zhao et al., 2005). In our model, we anticipate that much of this effect will be channeled through attitude and PBC – as AI self-efficacy makes one more optimistic and confident, which then increases intention – but a residual direct link is plausible. Accordingly, the following hypotheses are formulated.

H8: AI self-efficacy is positively correlated with a) attitude towards sustainability-oriented entrepreneurship, (b) perceived

behavioral control, and (c) sustainability-oriented entrepreneurial intentions.

H9: AI self-efficacy is indirectly correlated with sustainability-oriented entrepreneurial intentions via (a) attitude towards

sustainability-oriented entrepreneurship and (b) perceived behavioral control.

The conceptual framework is summarized in Figure 1.

Figure 1. Hypothesized model

METHODOLOGY

Research sample

Undergraduate students were selected as the research sample for several reasons. Prior studies suggest that university students are a particularly suitable population for examining entrepreneurial intentions, as they are typically at a career decision stage where entrepreneurship is a viable and actively considered option (Adam et al., 2025; Al Issa et al., 2025). University students tend to be more homogeneous in terms of prior entrepreneurial experience compared to the general adult population, which reduces potential biases and enhances internal validity (Ahmed et al., 2025). Furthermore, the dynamic economic environment in Vietnam, which has shifted from an anti-entrepreneurial to a pro-entrepreneurial orientation (Nguyen, 2023), provides an ideal backdrop for studying entrepreneurial aspirations. In addition, recent evidence suggests that nearly 90% of university students have actively engaged with AI tools and GenAI platforms, such as ChatGPT, BingAI, Grammarly, and DeepSeek, for their academic learning activities (Pan et al., 2025). This widespread exposure to AI technology among university students further justifies their selection, as it aligns closely with the study’s focus on AI knowledge, AI self-efficacy, and sustainability-oriented entrepreneurial intentions.

A purposive cluster sampling approach with gatekeeper selection was employed. In the first stage, the selection process was concentrated in two major regions of Vietnam, the Northern and Southern regions, where 224 higher education institutions are located (123 in the North and 101 in the South). Data were collected from undergraduate students at five universities in Vietnam between 20th February and 20th March 2025. In the first stage, two major regions of Vietnam, the Northern and Southern regions, were identified as the geographic frame. In the second stage, three universities in the North and two universities in the South were chosen based on their prominence in the Webometrics “impact” rankings (Webometrics, 2023), which served as a pragmatic stratification criterion. In the third stage, Student Affairs Officers and representatives from the Departments of Training Management acted as gatekeepers to identify students who met the inclusion criteria: those currently enrolled in, or who had completed, at least one entrepreneurship-related course. Lecturers and assistants then facilitated the distribution of questionnaires and provided clarification when needed. In the final stage, questionnaires were distributed directly to students. Participation was entirely voluntary, and informed consent was obtained from all respondents. Confidentiality and anonymity were strictly assured. Institutional approval for the survey procedures was obtained prior to data collection. However, given the purposive and convenience-based nature of the sampling strategy, the findings should be interpreted with caution. Claims of generalizability are limited to entrepreneurship-exposed university students in Vietnam and may not extend to the broader population of young people or entrepreneurs.

A total of 500 questionnaires were distributed, yielding 225 returned responses. After excluding eight incomplete questionnaires, a final sample of 217 valid responses was retained for further analysis. Regarding age distribution, most participants (77.4%) were between 20 and 23 years old, while 7.8% were aged 18 to 19 years, and 14.7% were older than 23 years. In our study, Age was coded as a categorical variable with three groups: 1 = 18–19 years, 2 = 20–23 years, and 3 => 23 years. In terms of gender, 46.1% of the respondents were male, and 53.9% were female. Gender was coded as a binary variable: 1 = Male, 2 = Female. Concerning their fields of study, 57.1% were enrolled in business-related disciplines, and 42.9% were pursuing non-business fields. ield was coded as 1 = Business-related, 2 = Non-business-related.

Scales

The measurement scales used in this study were adapted from validated instruments developed in prior research. All items were rated on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). AI knowledge was measured using four items from Chiu et al. (2024), while AI self-efficacy was assessed through four items adapted from Latikka et al. (2019). Attitude towards sustainability-oriented entrepreneurship and perceived behavioral control were each measured with four items based on Sharma, Bulsara, Bagdi, et al. (2023). Sustainability-oriented entrepreneurial intentions were captured through five items following the scale developed by Sharma, Bulsara, Bagdi, et al. (2023). The sources and detailed information for each construct are presented in Table 1.

Analytical approach

To examine the relationships proposed in the conceptual framework, we employed a multi-step analytical strategy. First, we conducted confirmatory factor analysis (CFA) to assess the validity and reliability of the measurement model, including the constructs’ internal consistency, convergent validity, and discriminant validity. CFA was chosen to ensure that each latent variable was accurately represented by its observed indicators, a prerequisite for meaningful hypothesis testing. After establishing measurement validity, we used polynomial regression with response surface analysis (PSA) to test the interaction effects between ATS and PBC on SOI. Polynomial regression and PSA enable the modeling of congruence and incongruence effects between two predictors, providing a nuanced understanding of their joint impact beyond traditional linear models (Anders et al., 2023). Following the guidance of Shanock et al. (2010), all variables in Model 3 of Table 3 were mean-centered before creating the squared and cross-product terms to improve interpretability and reduce multicollinearity. Variance inflation factors (VIFs) were computed for all predictors, and the values were below the conventional threshold of 5, confirming the absence of problematic collinearity. This approach is particularly well-suited for exploring how alignment (or misalignment) between cognitive factors influences entrepreneurial intentions.

To further explore direct and indirect relationships among the constructs, including the mediating roles of ATS and PBC, we employed the PROCESS macro (Model 4) for mediation analysis with bootstrap confidence intervals. Specifically, 5,000 bootstrap resamples were used to generate confidence intervals, ensuring robust estimation of indirect effects. In addition, we clarify that all PROCESS effects were estimated in an unstandardized form, consistent with Hayes’s (2018) guidelines. This technique enables the testing of multiple mediation paths in a statistically rigorous manner, providing insight into the mechanisms through which AI knowledge and AI self-efficacy influence SOI. Moreover, to ensure the robustness of our results, we additionally conducted structural equation modeling (SEM) using SmartPLS 4.0. The SEM approach allowed us to simultaneously estimate all hypothesized relationships and test the overall fit of the conceptual model. Furthermore, SmartPLS was used to assess the potential interaction effects between AI knowledge and AI self-efficacy, as well as to perform nested model comparisons for evaluating the robustness and parsimony of the findings. The consistency of results across these multiple analytical techniques reinforces the credibility of our conclusions.

Scale assessment

Confirmatory factor analysis (CFA) was conducted to assess the reliability and validity of the measurement model. One item from the AI self-efficacy construct (AIS4) was removed due to a low standardized loading of 0.402, which did not meet the minimum threshold of 0.50 (Brown, 2006). After eliminating the unsatisfactory item, the revised five-factor model demonstrated an acceptable fit to the data: χ²(204) = 292.295, p < 0.001; χ²/df = 1.433; GFI = 0.901; AGFI = 0.866; CFI = 0.968; TLI = 0.960; NFI = 0.902; and RMSEA = 0.045, SRMR = 0.0534, indicating good model fit according to conventional criteria (Hair et al., 2021) (see Figure 2). The CFA was estimated using maximum likelihood (ML), and the five-point Likert items were treated as ordinal but approximated as continuous, consistent with common practice when distributions do not severely deviate from normality.

Figure 2. Measurement model

As shown in Table 1, all constructs achieved satisfactory reliability, with α values exceeding 0.70 and CR values above 0.70. Convergent validity was supported as all constructs’ AVE exceeded the 0.50 threshold (Anderson & Gerbing, 1988), except for AIK. Although AVEAIK only accounted for 0.472, its CRAIK is 0.779, which is greater than 0.7; it is thus acceptable for further analyses (Brown, 2006). Discriminant validity was confirmed, as the square roots of the AVE values, presented along the diagonal of Table 2, were greater than the corresponding inter-construct correlations (Chin, 1998).

Common method bias

Harman’s single-factor test was conducted to assess potential common method bias (CMB). The unrotated first factor accounted for 37.675% of the total variance, which is below the recommended threshold of 50% (Harman, 1976), suggesting that common method bias is not a major concern. Second, we implemented a common latent factor (CLF) approach to more rigorously assess the possibility of method bias. In this model, a latent method factor was added to capture the variance shared across all items, following current best practices. The comparison of standardized factor loadings between the model with and without the CLF revealed that the differences were all below 0.20, which is within the acceptable range (Kock, 2021). This result indicates that common method variance did not substantially bias the estimates in our study. Finally, we also employed procedural remedies to mitigate the risk of CMB further. These included ensuring respondent anonymity, separating measurement of predictor and criterion variables in the questionnaire design, and randomizing item order to reduce consistency motifs.

RESULTS

Hypothesis testing

Multiple linear regression, polynomial regression, the PROCESS macro, and response surface analysis (RSA) were performed to examine the proposed hypotheses. The detailed results are presented in Tables 3 and 4 and visualized in Figure 3.

The findings first confirm the positive effects of ATS and PBC on SOI. As shown in Model 3 of Table 3, ATS (β = 0.361, p < 0.001) and PBC (β = 0.338, p < 0.001) were both significant predictors of SOI, supporting H1 and H2. In contrast, SN did not have a significant effect on SOI (β = 0.045, p = 0.489); thus, H3 was not supported. The response surface analysis revealed significant effects along the congruence line (ATS = PBC). The slope (ё₁) along this line was positive and significant (ё₁ = 0.700, p < 0.001), and importantly, the curvature (ё₂) was also significant (ё₂ = 0.150, p = 0.024), indicating a nonlinear relationship. This suggests that SOI increases more rapidly when both ATS and PBC are simultaneously high. This finding supports H4, confirming that SOI reaches its peak when ATS and PBC are aligned and high. As illustrated in the response surface plot (see Figure 3), the highest levels of SOI are observed at the back corner of the graph, where both ATS and PBC are high, while the lowest levels are found where both variables are low. In contrast, the analysis of the incongruence line (ATS ≠ PBC) indicated that neither the slope (ё₃ = 0.020, p = 0.879) nor the curvature (ё₄ = 0.330, p = 0.133) was significant. These results imply that discrepancies between ATS and PBC do not significantly diminish SOI, thus failing to support H5. The response surface plot demonstrates that SOI remains relatively stable even when ATS and PBC are misaligned, suggesting that individuals may maintain entrepreneurial intentions despite a moderate imbalance between attitude and perceived behavioral control.

The direct effects of AIK and AIS on ATS and PBC were also significant. As indicated in Model 1 and Model 2 of Table 3, AIK positively influenced ATS (β = 0.341, p < 0.001) and PBC (β = 0.306, p < 0.001), while AIS also positively influenced ATS (β = 0.234, p < 0.001) and PBC (β = 0.371, p < 0.001), supporting H6a, H6b, H8a, and H8b. However, neither AIK (β = 0.053, p = 0.513) nor AIS (β = 0.082, p = 0.196) had a significant direct effect on SOI, failing to support H6c and H8c. Mediation effects were further examined using bootstrap analyses, as reported in Table 4. The indirect effects of AIK on SOI through ATS (β = 0.189, 95% CI [0.081, 0.308]) and through PBC (β = 0.142, 95% CI [0.056, 0.247]) were both significant, supporting H7a and H7b. Similarly, AIS exhibited significant indirect effects on SOI through ATS (β = 0.137, 95% CI [0.060, 0.225]) and through PBC (β = 0.120, 95% CI [0.042, 0.211]), supporting H9a and H9b.

Table 1. Internal consistency and convergent validity

Constructs/Scales	Codes	Measures	Sources	Loading	α	CR	AVE
AI knowledge	AIK1	I can distinguish whether a tool is AI-based or not	Chiu et al. (2024)	0.604	0.798	0.779	0.472
	AIK2	I can create content with AI		0.581
	AIK3	I can explain what AI is		0.790
	AIK4	I know how to choose the right AI tools to effectively complete a task		0.750
AI self-efficacy	AIS1	I feel confident in understanding how AI caregiving robots works	Latikka et al. (2019)	0.829	0.864	0.870	0.690
	AIS2	I am confident in my decision-making regarding the use of AI caregiving robots		0.873
	AIS3	I possess the necessary skills to effectively use AI caregiving robots		0.788
	AIS4	I am confident in my understanding of the benefits and value that AI caregiving robots offer		-
Attitude towards sustainability-oriented entrepreneurship	ATS1	For me, a career as a sustainable entrepreneur is attractive	Sharma, Bulsara, Bagdi, et al. (2023)	0.689	0.806	0.826	0.547
	ATS2	Rather than working for a company, I would rather be a sustainable entrepreneur		0.847
	ATS3	I am positive about becoming a sustainable entrepreneur		0.806
	ATS4	Being a sustainable entrepreneur, in my opinion, is quite desirable		0.590
Perceived behavioral control	PBC1	It would be easy for me to start and run a sustainable enterprise	Sharma, Bulsara, Bagdi, et al. (2023)	0.772	0.842	0.837	0.562
	PBC2	I am confident in my ability to identify new business opportunities		0.718
	PBC3	I think I have sufficient traits to start a sustainable firm		0.744
	PBC4	I would have a good chance of succeeding if I tried to start a company		0.763
Sustainability-oriented entrepreneurial intentions	SOI1	I wished to start a sustainability-oriented enterprise that assists in alleviating environmental issues during my study at the university	Yi (2020)	0.796	0.890	0.889	0.618
	SOI2	I had a preliminary idea for a sustainability-oriented enterprise to implement in the future during my studies at university.		0.859
	SOI3	My professional goal was to become a sustainability-oriented entrepreneur during my studies at the university		0.631
	SOI4	I was willing to do anything to become a sustainability-oriented entrepreneur during my studies at the university		0.816
	SOI5	I would act as a professional manager and get involved in the management of a sustainability-oriented enterprise through promotion and preparation during my studies at the university.		0.809
Subjective norms	SN1	My close family would approve of the decision to create a sustainable business	Truong et al. (2022)	0.853	0.876	0.880	0.709
	SN2	My friends would approve of the decision to create a sustainable business		0.863
	SN3	My colleagues would approve of the decision to create a sustainable business		0.810

Note: N = 217, α: Cronbach’s alpha; CR = Composite reliability; AVE = Average variance extracted.

Table 2. Descriptive statistics and correlations

Variables	Mean	S.D.	1	2	3	4	5	6	7	8	9
Gender	1.539	0.500	-
Age	2.069	0.471	0.038	-
Fields	1.429	0.496	0.072	0.071	-
SOI	3.063	0.952	-0.036	0.106	-0.024	0.786
AIK	3.618	0.776	0.003	0.013	0.019	0.357**	0.687
AIS	3.121	1.008	-0.078	0.148*	0.007	0.405**	0.363**	0.831
ATS	3.501	0.745	0.023	0.062	-0.030	0.530**	0.469**	0.443**	0.740
PBC	3.230	0.920	0.043	0.115	0.018	0.506**	0.407**	0.502**	0.568**	0.750
SN	3.124	1.020	0.016	0.107	-0.030	0.363**	0.409**	0.377**	0.354**	0.538**	0.842

Note: N = 217. *p < 0.05, ** p < 0.01. The bolded scores along the diagonal represent the square root of the AVE.

Table 3. Hypothesis testing

Variables	Attitude towards sus. entrepreneurship				Perceived behavioural control				Sustainable entrepreneurial intentions
	Model 1				Model 2				Model 3
	β	SE	t	p-value	β	SE	t	p-value	β	SE	t	p-value
Constant	1.476***	0.320	4.608	<0.001	0.555	0.393	1.411	0.160	-0.142	0.312	-0.457	0.648
Age	0.019	0.092	0.211	0.833	0.096	0.113	0.848	0.397	0.098	0.112	0.876	0.382
Gender	0.074	0.086	0.861	0.390	0.133	0.106	1.262	0.208	-0.116	0.105	-1.104	0.271
Fields of study	-0.064	0.086	-0.744	0.458	0.004	0.106	0.039	0.969	-0.017	0.106	-0.165	0.869
AIK	0.341***	0.059	5.766	<0.001	0.306***	0.073	4.224	<0.001	0.053	0.080	0.656	0.513
AIS	0.234***	0.046	5.081	<0.001	0.371***	0.057	6.552	<0.001	0.082	0.063	1.298	0.196
SN									0.045	0.064	0.693	0.489
θ1: ATS 0.361*** 0.094											3.849	<0.001
θ2: PBC									0.338***	0.089	3.789	<0.001
θ3: ATS2									0.077	0.099	0.775	0.439
θ4: ATS x PBC									-0.092	0.109	-0.845	0.399
θ5: PBC2									0.162*	0.067	2.415	0.017
R2	0.310				0.318				0.393
Adjusted R2	0.291				0.302				0.360
F Change	18.971***				19.664***				12.041***
Congruence line (ATS = PBC)
ё1: Slope (θ1 + θ2)									0.700***	0.100	6.740	<0.001
ё2: Curvature (θ3 + θ4 + θ5)									0.150*	0.060	2.276	0.024
Incongruence line (ATS = -PBC)
ё3: Slope (θ1 - θ2)									0.020	0.150	0.152	0.879
ё4: Curvature (θ3 - θ4 + θ5)									0.330	0.220	1.508	0.133

Note: N = 217, *p < 0.05. **p < 0.01, ** p < 0.001.

Figure 3. Response surface analysis of ATS and PBC

Table 4. Indirect effect analyses

Indirect paths					Effects	BootSE	Bootstrap 95% CIs
								LLCI	ULCI
AIK	→	ATS	→	SOI	0.189	0.058		0.081	0.308
AIK	→	PBC	→	SOI	0.142	0.048		0.056	0.247
AIS	→	ATS	→	SOI	0.137	0.042		0.060	0.225
AIS	→	PBC	→	SOI	0.120	0.043		0.042	0.211

Note: N = 217; LLCI: Lower level of confidence interval; ULCI: Upper level of confidence interval; SE: Standard errors.

Robustness test

To address concerns regarding the analytical approach and potential limitations of the polynomial regression and response surface analysis, we conducted additional robustness checks using structural equation modeling (SEM) with SmartPLS 4.0. This method enabled us to simultaneously assess multiple mediation, moderation, and interaction effects, providing a comprehensive validation of our original findings.

The SEM results, as shown in Table 5 and Figure 4, closely mirror those obtained from the main analyses. Both AIK and AIS maintained significant positive effects on ATS and PBC. In turn, ATS and PBC remained significant predictors of SOI, whereas the direct effects of AIK and AIS on SOI were not significant, again confirming full mediation through TPB pathways.

Importantly, we tested the interaction effects between AIK and AIS on all key outcomes, as recommended by the reviewer. Across all tested paths, the interaction terms (AIK×AIS and AIS×AIK) did not reach statistical significance (all p > 0.05), indicating that the synergy between AI knowledge and AI self-efficacy does not exert an additional effect on SOI beyond their individual contributions. These results further corroborate our initial conclusion that AIK and AIS primarily operate independently via attitudes and perceived control.

Additionally, the inclusion of SN in the model did not alter the pattern of results, as SN remained a non-significant predictor of SOI. Overall, the robustness analyses using SEM confirm the validity and stability of our main findings. The lack of significant interaction effects suggests that AI-related competencies individually, rather than synergistically, shape the cognitive antecedents of sustainability-oriented entrepreneurial intentions in our sample. This multi-method validation increases confidence in our results and addresses concerns about methodological limitations.

Table 5. Robustness test

Paths					β	S.D.	t-value	p-value	95% CIs
									LL	UL
AIK	→	ATS			0.362 ***	0.065	5.565	0.000	0.230	0.487
AIK	→	PBC			0.272 ***	0.065	4.152	0.000	0.148	0.405
AIK	→	SOI			0.043	0.074	0.587	0.557	-0.101	0.186
AIS	→	ATS			0.308 ***	0.064	4.813	0.000	0.182	0.431
AIS	→	PBC			0.403 ***	0.075	5.389	0.000	0.251	0.544
AIS	→	SOI			0.109	0.072	1.508	0.132	-0.029	0.253
ATS	→	SOI			0.302 **	0.089	3.386	0.001	0.120	0.473
PBC	→	SOI			0.243 **	0.090	2.701	0.007	0.060	0.415
SN	→	SOI			0.070	0.085	0.814	0.416	-0.091	0.242
AIS*AIK	→	ATS			0.051	0.058	0.872	0.383	-0.067	0.163
AIS*AIK	→	PBC	→	SOI	0.001	0.018	0.033	0.973	-0.033	0.041
AIK	→	ATS	→	SOI	0.109 **	0.039	2.778	0.005	0.039	0.192
AIS	→	ATS	→	SOI	0.093 **	0.033	2.816	0.005	0.033	0.163
AIS	→	PBC	→	SOI	0.098 *	0.039	2.502	0.012	0.025	0.179
AIK	→	PBC	→	SOI	0.066 *	0.030	2.214	0.027	0.015	0.134
AIS*AIK	→	ATS	→	SOI	0.015	0.019	0.787	0.432	-0.020	0.058

Note: N= 217. *p < 0.05. **p < 0.01. ***p < 0.001; S.D.: Standard deviation; CIs: Confidence intervals; LL: Low limit (2.5%); UL: Upper limit (97.5%).

Figure 4. Robustness test

DISCUSSIONS

This study set out to investigate how AI-related competencies (knowledge and self-efficacy) shape sustainability-oriented entrepreneurial intentions through the cognitive pathways of attitude and perceived behavioral control, as grounded in the TPB. Overall, our findings provide strong support for the hypothesized model, with a few notable nuances.

First, consistent with TPB and prior entrepreneurship research, we found that both attitude towards SOE and perceived behavioral control had significant positive effects on sustainability-oriented entrepreneurial intentions. It means that students who held a more favorable attitude about engaging in sustainable entrepreneurship and those who felt a greater sense of control over launching a venture were more likely to express intentions to start a sustainability-oriented business. This aligns with numerous studies emphasizing attitude and self-efficacy as critical drivers of entrepreneurial intentions (Amani et al., 2024; Sharma, Bulsara, Bagdi, et al., 2023). Our results reaffirm that even in the specific domain of sustainability, where the venture’s goals extend beyond profit, these personal motivational factors remain paramount. If an individual does not believe in the value of sustainable entrepreneurship or does not feel capable of performing it, they are unlikely to form an intention to pursue it. These findings align with those of Vuorio et al. (2017) and Tan et al. (2020), who also reported positive contributions of attitude and perceived feasibility to sustainable or social entrepreneurial intentions. We thereby extend their observations to a novel context involving AI considerations.

Turning to the interplay between attitude and perceived behavioral control, the response surface analysis provides important insights. We hypothesized that congruence between high attitude and high PBC would enhance intentions, while incongruence would reduce them. The findings partially supported these predictions. Intentions were highest when both attitude and PBC were strongly favorable, as indicated by the significant positive slope along the line of congruence (attitude = PBC). This suggests a synergistic effect: individuals who are both highly motivated towards sustainable entrepreneurship and confident in their ability to act are most committed to entrepreneurial action. This outcome also aligns with Shapero’s model, which emphasizes the joint importance of desirability and feasibility (Krueger et al., 2000). Furthermore, we found a significant positive curvature along the line of congruence, indicating a convex relationship. This suggests that the benefit of alignment becomes even more pronounced when both attitude and PBC are simultaneously high, providing additional support for the additive mechanism.

However, no significant punitive effect of incongruence was found. An imbalance between attitude and PBC did not significantly lower intentions compared to individual levels alone. The surface analysis showed that individuals with mismatched levels of attitude and PBC maintained moderate intention levels, suggesting an additive rather than multiplicative relationship. One strong factor appeared to partially compensate for the weaker one: high passion could sustain intentions despite low perceived control, and strong confidence could sustain intentions despite weaker sustainability commitment. This finding indicates that strengthening either attitude or PBC could independently foster sustainability-oriented entrepreneurial intentions. Although these results diverge from some theoretical expectations, they are consistent with empirical findings that found limited interaction effects between TPB components (Ahmed et al., 2025; Relente & Capistrano, 2024). In contrast to expectations based on the canonical TPB model, subjective norms did not exhibit a statistically significant effect on SOI. This finding suggests that perceived social approval from family, friends, and colleagues may not be a key determinant of students’ intention to pursue sustainability-oriented entrepreneurship. This could reflect a broader trend observed in prior studies, where subjective norms were found to be weaker predictors of entrepreneurial intention compared to personal attitudes and perceived control (Heredia-Carroza et al., 2024; Sharma, Bulsara, Trivedi, et al., 2023). Another possible explanation lies in the growing autonomy of career decision-making among younger generations, who may prioritize internal motivations over external social expectations (Tran et al., 2023). Alternatively, in the specific Vietnamese university context studied, students may not yet perceive strong societal endorsement or peer pressure toward sustainable entrepreneurship (Truong et al., 2022).

A central contribution of this research is to illuminate the role of AI competencies in fostering sustainable entrepreneurial intentions. The data robustly supported our hypotheses regarding AI knowledge and AI self-efficacy, with some interesting patterns. We found that AI knowledge had a positive influence on both attitude towards SOE and perceived behavioral control. Students who possessed greater knowledge about AI tended to evaluate sustainable entrepreneurship more positively and felt more capable of starting such a venture. This result resonates with the argument that understanding technology can make sustainable business ideas seem more feasible and attractive (Bickley et al., 2025; Gupta et al., 2023). It appears that AI-literate individuals are more likely to see the potential of integrating AI solutions for sustainability challenges, which boosts their enthusiasm (attitude) for these ventures. Simultaneously, their familiarity with AI tools gives them confidence (PBC) that they can implement the technical side of a sustainable enterprise. These relationships mirror findings in general entrepreneurship education literature, where knowledge acquisition enhances both perceived desirability and feasibility of entrepreneurship (Fayolle & Gailly, 2015).

However, neither AI knowledge nor AI self-efficacy had a significant direct effect on SOE intentions when attitude and PBC were taken into account. This indicates that the influence of AI competencies on intentions is fully mediated by attitude and PBC, aligning with the TPB notion that such external factors work through the proximal determinants (Ajzen, 1991). In other words, simply knowing a lot about AI does not automatically translate into wanting to start a sustainable business – it translates into wanting and feeling able to do so, which then drives the intention. This full mediation is a significant theoretical confirmation, underscoring the value of the TPB framework in explaining how background factors, such as technological knowledge, influence entrepreneurial motivation. It also cautions that boosting AI knowledge alone may not increase entrepreneurial intentions unless it alters how people perceive the venture and their own capabilities.

With respect to AI self-efficacy, we observed similar patterns. Higher AI self-efficacy was associated with more positive attitudes towards SOE and higher PBC. This suggests that confidence in using AI engenders a mindset conducive to sustainable entrepreneurship. Those students who believed in their ability to work with AI were more optimistic about starting a venture that likely involves technology (improving their attitude) and felt fewer hurdles in doing so (elevating their perceived control). These findings are in line with broader entrepreneurship studies highlighting self-efficacy as a potent antecedent of attitudes and perceived feasibility (Newman et al., 2019). Our results extend this knowledge to the specific case of AI-related self-efficacy, showing that such domain-specific confidence can spill over into the entrepreneurial context. Again, the direct path from AI self-efficacy to intention was not significant, reinforcing that its effect on intentions is channeled via attitude and PBC. This full mediation was evidenced by the significant indirect effects: AI self-efficacy increased SOE intentions through its positive impact on attitude and on PBC. Essentially, an individual who feels adept with AI will likely think more positively about a sustainable venture and feel in control, which in turn makes them more inclined to pursue that venture. This insight echoes prior research on self-efficacy in green entrepreneurship – for example, recent work by Al Issa et al. (2025) found that entrepreneurial self-efficacy (bolstered through experiential learning) significantly raised social entrepreneurial intentions by influencing individuals’ confidence and aspirations.

Nonetheless, while these findings position AI-related competencies as important enablers, we caution that the adoption of AI in entrepreneurship is not without critical challenges. First, the transformative potential of AI for sustainability is accompanied by significant risks and contradictions. Many AI tools are embedded within capital- and resource-intensive infrastructures and may contribute to increased energy use and environmental burdens (Wang et al., 2025). This “digital sustainability paradox” suggests that while AI can advance sustainable outcomes, its widespread adoption could also exacerbate ecological footprints if not properly managed. Furthermore, the integration of AI into entrepreneurship can widen digital divides, as access to advanced AI technologies and training is unevenly distributed—potentially deepening inequalities among aspiring entrepreneurs, especially across regions or socio-economic groups (Vinuesa et al., 2020). Equity in access to AI tools and knowledge remains a pressing issue for educators and policymakers. Second, we acknowledge that our study primarily focuses on the enabling side of AI competencies, without directly investigating or controlling for the broader socio-technical and ethical risks associated with these competencies. Future work should critically examine both the positive and negative implications of digitalization for sustainable entrepreneurship, addressing the systemic and institutional context in which technological adoption unfolds. Finally, while we modeled and discussed subjective norms, further research should consider their indirect and contextual effects—especially in collectivist societies—where social and institutional support or resistance can shape entrepreneurial pathways.

CONCLUSIONS

Theoretical contributions

This research makes several important theoretical contributions. First, it extends the TPB into the novel intersection of AI and sustainability-oriented entrepreneurship. While TPB has been widely used in entrepreneurial intention studies, our study is among the first to incorporate AI-related personal factors into the TPB framework for sustainable entrepreneurship. By demonstrating that AI knowledge and AI self-efficacy have a significant influence on attitude and PBC, which in turn drive intentions, we provide evidence that the TPB can flexibly accommodate emerging determinants relevant to the digital age. This extension answers recent calls in the literature to integrate technological competencies into models of entrepreneurial intentions (Bui & Duong, 2024). We demonstrate that individuals’ technological readiness (in terms of knowledge and confidence) is an important part of the cognitive equation for whether they decide to embark on a sustainable venture. In doing so, we also reinforce Ajzen’s principle that background factors operate through proximal predictors: both AI knowledge and self-efficacy influence attitudes and control beliefs, rather than acting directly, thereby bolstering the theoretical premise of mediation in TPB. This contributes to the TPB literature by empirically confirming the mediated nature of external influences in a new context.

Second, our study contributes to the sustainable entrepreneurship theory by identifying concrete cognitive mechanisms through which the oft-discussed “technological enablers” affect sustainable entrepreneurial intentions. Prior research has acknowledged that technology (like digital platforms or green tech) can empower sustainable entrepreneurship (Lourenço et al., 2024), but there has been limited understanding of how this happens at the individual psychological level. Our findings suggest that simply having access to advanced technologies like AI is not enough – what matters is how these technologies are perceived and understood by the potential entrepreneur. We show that when individuals have internalized AI skills and confidence, they are more likely to view sustainable ventures as desirable and feasible, thereby increasing their intent to pursue them. This provides a more nuanced theoretical insight: the impact of technological advancements on sustainable entrepreneurship is mediated by human capital and cognitive perceptions. It bridges the gap between macro-level discussions of technology trends and micro-level analyses of entrepreneurial intention formation. This insight can be incorporated into sustainable entrepreneurship models, emphasizing personal technology readiness as part of the entrepreneurial “toolkit” that shapes venture creation decisions.

Third, our research makes a methodological and conceptual contribution by examining the interaction of attitude and perceived behavioral control using polynomial regression and response surface analysis. Entrepreneurship scholars have long debated whether the components of intention models simply add up or whether they have interactive effects (Schlaegel & Koenig, 2014). We introduced and tested the concept of congruence versus incongruence between attitude and PBC in determining intention. While our hypothesis of a strong diminishment of intention under incongruence was not supported, the approach itself yielded rich information. We found evidence of a primarily additive relationship, implying that TPB’s linear additive assumption held in our data – but we also confirmed that the “best case” scenario for intention is when both attitude and PBC are high (the ridge of the response surface). The lack of significant depression, despite a misaligned attitude/PBC, adds a new perspective: it suggests a certain robustness or compensatory effect in entrepreneurial motivation. This is a theoretical nuance that complements TPB: even if one cognitive driver is suboptimal, the other can buffer the effect to some degree. Future theoretical models of entrepreneurial intention could consider incorporating threshold or minimum conditions (e.g., perhaps one of the two must be above a certain level) rather than a strict interaction term. Our study thus contributes to the fine-tuning of intention theory in entrepreneurship, and it demonstrates the utility of advanced analytical techniques in unpacking these relationships beyond simple linear regression.

Finally, we enrich the literature on entrepreneurship in the digital era by combining insights from entrepreneurship, sustainability, and information systems. Our integrative approach shows that theories from one domain (like TPB from social psychology) can successfully interface with constructs from another domain (AI competencies from technology/education literature) to explain a complex phenomenon (sustainable entrepreneurship intention). The positive results lend support to an interdisciplinary theoretical outlook: understanding modern entrepreneurship requires considering technological factors as part of individuals’ cognitive frameworks. In doing so, we provide a foundation for future research to explore other technology-related variables (such as actual usage of AI tools or attitudes specifically toward technology) within entrepreneurship theories. We also contribute to the emerging discourse on “sustainable digital entrepreneurship,” positioning our work at the convergence of digital transformation and sustainability in entrepreneurship. Theoretically, this helps build a narrative that sustainable entrepreneurs of the future will be those who can effectively harness digital innovations, and our study offers a tested model of how that convergence plays out at the intention stage.

Practical implications

Our findings carry important practical implications for educators, trainers, policymakers, and prospective entrepreneurs, especially in university and startup ecosystem settings.

Firstly, the significant role of attitude and perceived behavioral control in driving SOE intentions suggests that interventions aiming to foster sustainable entrepreneurship should focus on enhancing these cognitive factors. Educational programs in universities can help build positive attitudes towards sustainability-oriented businesses by showcasing success stories of impactful sustainable startups, incorporating sustainability challenges into entrepreneurship courses, and facilitating value-driven reflection among students. If students come to genuinely appreciate the societal and environmental value of sustainable ventures, their attitudes are likely to become more favorable. Simultaneously, programs should aim to increase students’ perceived behavioral control. This can be achieved by equipping them with relevant skills (through workshops on business planning and sustainable innovation, for example), providing mentorship, and even conducting simulations of venture creation to allow them to practice entrepreneurial tasks. The more competent and in control students feel, the more likely they are to translate their interest into intention. Given that our results show intentions are maximized when both attitude and PBC are high, a balanced approach addressing both mindset and skillset is advisable. Given the additive effects identified in our response surface analysis, practitioners should seek a balanced approach that simultaneously builds both mindset and skillset, as the highest intentions were observed when both were high.

A second implication revolves around integrating AI and digital competencies into entrepreneurship development for sustainability. We found that AI knowledge and AI self-efficacy indirectly boosted intentions via attitude and PBC. This means that simply teaching about sustainability or business might not be enough; integrating technology training is key. Universities and incubators should consider adding AI-oriented modules in their entrepreneurship curriculum – for example, introducing students to AI tools that can be used in sustainable ventures (like data analytics for climate data, AI in recycling processes, or machine learning for social impact measurement). By doing so, students not only gain knowledge but can also practice with these tools, building their confidence (self-efficacy). Our research suggests that when aspiring entrepreneurs feel competent with AI, they will also feel more capable of launching a business that uses those tools, thereby indirectly encouraging entrepreneurial action. Importantly, this implies a synergy: promoting “tech for good” skill-building can serve the dual purpose of improving technical literacy and spurring sustainable entrepreneurship. For entrepreneurship support organizations, offering training sessions on AI for sustainability or hackathons that blend AI and environmental problem-solving could ignite interest and self-belief among participants, leading to more startup ideas in this space. Importantly, access to such training should be broadened to avoid reinforcing digital divides, and educators should also address the ethical, environmental, and infrastructural challenges associated with AI adoption. By doing so, students not only gain knowledge but can also practice with these tools, building their confidence (self-efficacy).

Policymakers and innovation funding bodies can draw on our findings to design initiatives that encourage sustainability-oriented startups. One implication is to support the creation of interdisciplinary teams or programs that bring together students from computer science (with AI expertise) and students from environmental or business fields. Our results show that both knowledge and confidence in AI are assets – so if an individual entrepreneur lacks them, having a co-founder or team member with those strengths could compensate, boosting the team’s overall attitude and perceived control. Policymakers could fund innovation labs or accelerators that explicitly pair tech-savvy youth with sustainability-focused youth to work on venture projects. Additionally, because attitude (i.e., personal commitment to sustainability) remains crucial, programs like innovation competitions or startup grants could put weight on sustainability impact, thereby attracting those who are passionate about these issues and reinforcing the societal importance (which bolsters positive attitude across the community of entrepreneurs).

Finally, our finding that misalignment between attitude and PBC did not drastically impede intentions suggests that practitioners can take a strength-based approach. If a prospective entrepreneur clearly has one strong suit (either high passion or high capability), mentors can work to leverage that strength while gradually addressing the weaker area. For instance, an idealistic student who lacks confidence might be encouraged to pursue their idea with the reassurance of resources and support networks (thus gradually improving their PBC). Conversely, a very confident student who has not yet embraced the sustainability mission might be exposed to compelling evidence of impact and market demand in sustainability (to elevate their attitude). In practice, this means that entrepreneurship support should be personalized: understanding whether a given individual needs more motivation or more skill empowerment, and tailoring support accordingly, can help maintain their intention to start a sustainable venture. In conclusion, our study provides actionable insights that integrating AI competency development with sustainability entrepreneurship programs will likely yield more robust intentions and, eventually, more startups that contribute to sustainable development.

Limitations and suggestions for future studies

Despite its contributions, this study has several limitations that suggest directions for future research. First, our focus on Vietnamese undergraduate students limits the generalizability of findings. Cognitive and motivational dynamics may differ in other age groups, life stages, and national contexts. We acknowledge that homogeneity in educational attainment, life stage, and the national context constrains external validity. Future research should replicate the study in diverse cultural and professional settings, and consider comparative or subgroup analyses, such as those between business and non-business students, to explore potential moderating effects.

Second, although we conducted additional robustness checks using structural equation modeling (SEM)—including nested model comparisons and interaction effects between AI knowledge and AI self-efficacy—the relatively small sample size may have limited statistical power. The lack of significant interaction effects may result from type II error rather than the absence of a true relationship. Future studies should employ larger, more heterogeneous samples to validate our findings and fully exploit the potential of advanced analytical techniques, such as SEM for simultaneous mediation, moderation, and interaction testing.

Third, our measurement of AI knowledge and AI self-efficacy, although statistically discriminant, may still be affected by conceptual overlap and measurement adaptation. In particular, the AI self-efficacy scale was adopted from Latikka et al. (2019), which was originally developed in the context of AI caregiving robots. Although we respected the original validated scale for methodological rigor, this creates a domain restriction: the items may not fully capture entrepreneurship-relevant AI competencies, such as AI-based decision support, data analytics, or venture planning. The inclusion of a dropped item (AIS4, loading = 0.402) further indicates residual domain misfit. Consequently, the findings regarding AI self-efficacy should be interpreted with caution, as they may not generalize to broader AI applications in entrepreneurship. Future research should either adapt the wording of items to reflect more generic AI usage or develop new instruments tailored specifically to entrepreneurial contexts, thereby improving construct validity and enhancing the applicability of findings across domains.

Fourth, while we included subjective norms in our revised conceptual model to enhance theoretical fidelity, our operationalization of this construct was limited. Future studies should develop more nuanced measures of social influence, network effects, and collective norms, particularly in collectivist contexts where social pressures may play a significant role in shaping intentions. Further exploration of external, institutional, and policy-related factors—such as industry standards or government incentives—is also warranted.

Fifth, the theoretical framework of this study is primarily grounded in individual-level psychological constructs from the TPB, but it does not sufficiently capture the broader institutional, structural, and socio-technical barriers to sustainability-oriented entrepreneurship. We recommend that future research complement the TPB with perspectives from institutional theory, practice theory, or socio-technical transitions literature to investigate how policy, market structures, and collective action shape the relationship between intention and sustainable entrepreneurial behavior.

Moreover, the study employs a cross-sectional and intention-based design, capturing respondents’ perceptions and intentions at a single point in time rather than actual entrepreneurial behaviors. This approach limits the ability to infer causal relationships and predict whether reported intentions will translate into actual entrepreneurial action. Future research should employ longitudinal designs to track changes in AI knowledge, self-efficacy, and entrepreneurial intentions over time, as well as to examine whether and how these intentions ultimately result in the founding and performance of sustainability-oriented ventures.

Finally, while our findings confirm the mediating role of attitude and perceived behavioral control in the relationship between AI-related competencies and sustainability-oriented entrepreneurial intentions, future research should further test the robustness of full versus partial mediation using additional model comparison techniques, such as Sobel tests. As technological landscapes evolve, future research could also extend this framework to investigate the entrepreneurial implications of other digital innovations, such as blockchain or the Internet of Things, and critically engage with both the enabling potential and the risks of digital technologies. This includes ethical, environmental, and social concerns, as well as equity of access, ensuring that technological competencies, such as AI self-efficacy, are understood not only as enablers but also as constructs shaped by contextual limitations.

References

Aboobaker, N., D, R., & K.A, Z. (2023). Fostering entrepreneurial mindsets: The impact of learning motivation, personal innovativeness, technological self-efficacy, and human capital on entrepreneurial intention. Journal of International Education in Business, 16 (3), 312-333. https://doi.org/10.1108/jieb-10-2022-0071

Abu Shriha, S., Al-Shboul, M. d. A., & Abaddi, S. (2024). The e-entrepreneurial intentions of Jordanian business students to start an online business in emerging economies: an application of planned behavior theory. Management & Sustainability: An Arab Review 4(2), 237-267. https://doi.org/10.1108/msar-08-2023-0042

Adam, S., Mohd Fuzi, N., Ramdan, M. R., & Ismail, A. F. M. F. (2025). The effectiveness of digital entrepreneurship ecosystem toward enriching income generation: The moderating role of entrepreneurial intention. SAGE Open, 15(1), 1-16. https://doi.org/10.1177/21582440241305361

Agu, A. G., Kalu, O. O., Esi-Ubani, C. O., & Agu, P. C. (2021). Drivers of sustainable entrepreneurial intentions among university students: an integrated model from a developing world context. International Journal of Sustainability in Higher Education, 22(3), 659-680. https://doi.org/10.1108/ijshe-07-2020-0277

Ahmed, B. S., Wan Nawang, W. R., Saadallah, O., & Mursidi, A. (2025). Extending the theory of planned behavior in predicting entrepreneurial intention among university students: The role of perceived relational support. The International Journal of Management Education, 23(2), 101168. https://doi.org/10.1016/j.ijme.2025.101168

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–211. https://doi.org/10.1016/0749-5978(91)90020-T

Al-Romeedy, B. S., & El-Sisi, S. A.-W. (2024). Does AI have an effective role in applying sustainable entrepreneurship in the tourism industry? In D. Sharma, H. Abdullah, & P. Singh (Eds.), Sustainable tourism (Part A, pp. 3–22). Emerald Publishing. https://doi.org/10.1108/978-1-83797-979-020241001

Al Issa, H.-E., Thai, M. T. T., & Saad, S. (2025). Empowering social entrepreneurial intentions through experiential learning and self-efficacy. The International Journal of Management Education, 23(2), 101154. https://doi.org/10.1016/j.ijme.2025.101154

Alimehmeti, G., Ndoka, E., & Paletta, A. (2025). Cultivating green pioneers: examining the antecedents of sustainable entrepreneurial intent in higher education. International Journal of Sustainability in Higher Education. https://doi.org/10.1108/ijshe-03-2024-0206

Amani, D., Ismail, I. J., Makona, A., Changalima, I. A., & Kazungu, I. (2024). Extending the mediation role of entrepreneurial self-efficacy on enhancing students’ entrepreneurial intentions: A moderated mediation model. The International Journal of Management Education, 22(1), 1-12. https://doi.org/10.1016/j.ijme.2023.100915

Anders, C., Hooley, I., & Kivlighan, D. M. (2023). The nature of a pandemic: Testing the relationship between access to nature, nature relatedness, wellbeing and belonging in nature using polynomial regression with response surface analysis. Journal of Environmental Psychology, 85, 101949. https://doi.org/10.1016/j.jenvp.2022.101949

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. https://doi.org/https://doi.org/10.1037/0033-2909.103.3.411

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall.

Bellver, D. O., Pérez-Campos, C., González-Serrano, M. H., & Martínez-Rico, G. (2022). Towards the development of future sustainable sports entrepreneurs: An asymmetric approach of the sports sciences sustainable entrepreneurial intentions. Journal of Hospitality, Leisure, Sport & Tourism Education, 31, 100403. https://doi.org/10.1016/j.jhlste.2022.100403

Bewersdorff, A., Hornberger, M., Nerdel, C., & Schiff, D. S. (2025). AI advocates and cautious critics: How AI attitudes, AI interest, use of AI, and AI literacy build university students’ AI self-efficacy. Computers and Education: Artificial Intelligence, 8, 100340. https://doi.org/10.1016/j.caeai.2024.100340

Bickley, S. J., Macintyre, A., & Torgler, B. (2025). Artificial intelligence and big data in sustainable entrepreneurship. Journal of Economic Surveys, 39(1), 103–145. https://doi.org/10.1111/joes.12611

Bonfanti, A., De Crescenzo, V., Simeoni, F., & Loza Adaui, C. R. (2024). Convergences and divergences in sustainable entrepreneurship and social entrepreneurship research: A systematic review and research agenda. Journal of Business Research, 170, 114336. https://doi.org/10.1016/j.jbusres.2023.114336

Brown, T. A. (2006). Confirmatory factor analysis for applied research. Guilford Press.

Bui, H. N., & Duong, C. D. (2024). ChatGPT adoption in entrepreneurship and digital entrepreneurial intention: A moderated mediation model of technostress and digital entrepreneurial self-efficacy. Equilibrium. Quarterly Journal of Economics and Economic Policy, 19(2), 391-428. https://doi.org/10.24136/eq.3074

Chin, W. W. (1998). The partial least squares approach for structural equation modeling. In G. A. Marcoulides (Ed.), Modern methods for business research (pp. 295–336). Lawrence Erlbaum Associates.

Chiu, T. K. F., Ahmad, Z., & Çoban, M. (2024). Development and validation of teacher artificial intelligence (AI) competence self-efficacy (TAICS) scale. Education and Information Technologies, 30(5), 6667-6685. https://doi.org/10.1007/s10639-024-13094-z 

Duong, C. D. (2024). Exploring the role of cultural values on consumers’ organic food consumption: Does blockchain-enabled traceability matter? Oeconomia Copernicana, 15(4), 1509-1546. https://doi.org/10.24136/oc.3306

Duong, C. D. (2025). How AI‐enabled drivers inspire sustainability‐oriented entrepreneurial intentions: Unraveling the (in)congruent effects of perceived desirability and feasibility from the entrepreneurial event model perspective. Sustainable Development, 33 (4), 6228-6246. https://doi.org/10.1002/sd.3461

Duong, C. D., Bui, H. N., Chu, T. V., Van Pham, T., & Do, N. D. (2024). ICT skills, entrepreneurial self-perceived creativity, and digital entrepreneurship: Insights from the stimulus-organism-response model. Thinking Skills and Creativity, 54, 101646 https://doi.org/10.1016/j.tsc.2024.101646

Duong, C. D., & Nguyen, T. H. (2024). How ChatGPT adoption stimulates digital entrepreneurship: A stimulus-organism-response perspective. The International Journal of Management Education, 22(3), 101019. https://doi.org/https://doi.org/10.1016/j.ijme.2024.101019

Fayolle, A., & Gailly, B. (2015). The impact of entrepreneurship education on entrepreneurial attitudes and intention: Hysteresis and persistence. Journal of Small Business Management, 53(1), 75-93. https://doi.org/10.1111/jsbm.12065

Fazio, M. V., Freund, R., & Novella, R. (2024). Do entrepreneurial skills unlock opportunities for online freelancing? Experimental evidence from El Salvador. Journal of Development Economics, 172, 103363. https://doi.org/10.1016/j.jdeveco.2024.103363

Füller, J., Hutter, K., Wahl, J., Bilgram, V., & Tekic, Z. (2022). How AI revolutionizes innovation management: Perceptions and implementation preferences of AI-based innovators. Technological Forecasting and Social Change, 178, 121598. https://doi.org/10.1016/j.techfore.2022.121598

Füller, J., Hutter, K., Wahl, J., Bilgram, V., & Tekic, Z. (2022). How AI revolutionizes innovation management – Perceptions and implementation preferences of AI-based innovators. Technological Forecasting and Social Change, 178, 121598. https://doi.org/10.1016/j.techfore.2022.121598

Giuggioli, G., & Pellegrini, M. M. (2022). Artificial intelligence as an enabler for entrepreneurs: a systematic literature review and an agenda for future research. International Journal of Entrepreneurial Behavior & Research, 29(4), 816-837. https://doi.org/10.1108/ijebr-05-2021-0426

Guo, Q., Peng, Y., & Luo, K. (2025). The impact of artificial intelligence on energy–environmental performance: Empirical evidence from cities in China. Energy Economics, 141, 108136. https://doi.org/10.1016/j.eneco.2024.108136

Gupta, B. B., Gaurav, A., Panigrahi, P. K., & Arya, V. (2023). Analysis of artificial intelligence-based technologies and approaches on sustainable entrepreneurship. Technological Forecasting and Social Change, 186, 1-9. https://doi.org/10.1016/j.techfore.2022.122152

Hair, J. F., Hult, G. T. M., Ringle, C. M., Sarstedt, M., Danks, N. P., & Ray, S. (2021). Evaluation of reflective measurement models. In J. F. Hair Jr., G. T. M. Hult, C. M. Ringle, M. Sarstedt, N. P. Danks, & S. Ray (Eds.), Partial least squares structural equation modeling (PLS-SEM) using R: A workbook (pp. 75–90). Springer. https://doi.org/10.1007/978-3-030-80519-7_4

Hayes, A. F. (2018). An Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach (Third ed.). Guilford Press, New York.

Heredia-Carroza, J., Chavarría-Ortiz, C., López-Estrada, S., & Zacharewicz, T. (2024). How to enhance the entrepreneurial intentions of the young population in rural areas: An approach from personal values and the socioeconomic environment. European Research on Management and Business Economics, 30(3), 100261. https://doi.org/10.1016/j.iedeen.2024.100261

Hofstede, G. (2001). Culture’s Consequences. Comparing values, behaviors, institutions and organizations across nations (2nd ed.). Thousand Oaks, CA: Sage.

Hopp, C., Dey, P., Riniker, M., & Rüdiger, M. (2025). Financing sustainable entrepreneurship: Unpacking the role of campaign information and risk disclosure in reward-based crowdfunding. Finance Research Letters, 74, 106748. https://doi.org/https://doi.org/10.1016/j.frl.2025.106748

Hsu, D. K., Burmeister-Lamp, K., Simmons, S. A., Foo, M.-D., Hong, M. C., & Pipes, J. D. (2019). “I know I can, but I don’t fit”: Perceived fit, self-efficacy, and entrepreneurial intention. Journal of Business Venturing, 34(2), 311-326. https://doi.org/10.1016/j.jbusvent.2018.08.004

Imjai, N., Nui-Suk, C., Usman, B., Somwethee, P., & Aujirapongpan, S. (2024). The influence of AI competency and design thinking skills on innovative entrepreneurial competency: The role of strategic intelligence amongst new age entrepreneurs in Thailand. International Journal of Information Management Data Insights, 4(2), 100301. https://doi.org/https://doi.org/10.1016/j.jjimei.2024.100301

Ip, C. Y. (2024). From green entrepreneurial intention to behaviour: The role of environmental knowledge, subjective norms, and external institutional support. Sustainable Futures, 8, 100331. https://doi.org/10.1016/j.sftr.2024.100331

Kock, F., Berbekova, A., & Assaf, A. G. (2021). Understanding and managing the threat of common method bias: Detection, prevention and control. Tourism Management, 86, 104330. https://doi.org/10.1016/j.tourman.2021.104330

Krueger, N. F., Reilly, M. D., & Carsrud, A. L. (2000). Competing models of entrepreneurial intentions. Journal of Business Venturing, 15, 411–432.

Kuckertz, A., & Wagner, M. (2010). The influence of sustainability orientation on entrepreneurial intentions — Investigating the role of business experience. Journal of Business Venturing, 25(5), 524-539. https://doi.org/10.1016/j.jbusvent.2009.09.001

Kwilinski, A., Lyulyov, O., & Pimonenko, T. (2024). Green economic development and entrepreneurship transformation. Entrepreneurial Business and Economics Review, 12(4), 157-175. https://doi.org/10.15678/eber.2024.120409

Latikka, R., Turja, T., & Oksanen, A. (2019). Self-efficacy and acceptance of robots. Computers in Human Behavior, 93, 157–163. https://doi.org/10.1016/j.chb.2018.12.017

Lazarte-Aguirre, A. (2024). Pathways to sustainable entrepreneurship: Analysing drivers of sustainable entrepreneurial orientation. Sustainable Technology and Entrepreneurship, 3(3), 100081. https://doi.org/10.1016/j.stae.2024.100081

Lortie, J., & Castogiovanni, G. (2015). The theory of planned behavior in entrepreneurship research: what we know and future directions. International Entrepreneurship and Management Journal, 11(4), 935-957. https://doi.org/10.1007/s11365-015-0358-3

Lourenço, F., Lei, W. S., Couto, U., Lei, W. I., & Cheng, R. (2024). Driving sustainable entrepreneurship: Institutional and psychological influences at the Macao food festival. Tourism Management Perspectives, 53, 101298. https://doi.org/10.1016/j.tmp.2024.101298

Neneh, B. N. (2020). Entrepreneurial passion and entrepreneurial intention: the role of social support and entrepreneurial self-efficacy. Studies in Higher Education, 47(3), 587-603. https://doi.org/10.1080/03075079.2020.1770716

Newman, A., Obschonka, M., Schwarz, S., Cohen, M., & Nielsen, I. (2019). Entrepreneurial self-efficacy: A systematic review of the literature on its theoretical foundations, measurement, antecedents, and outcomes, and an agenda for future research. Journal of Vocational Behavior, 110, 403-419. https://doi.org/10.1016/j.jvb.2018.05.012

Nguyen, N. M. (2023). The effect of FDI on domestic entrepreneurship: the case of greenfield investment and cross-border M&A activities. Journal of Economics and Development, 25(1), 62-78. https://doi.org/10.1108/jed-11-2022-0228

Pan, M., Lai, C., & Guo, K. (2025). Effects of GenAI-empowered interactive support on university EFL students’ self-regulated strategy use and engagement in reading. The Internet and Higher Education, 65, 100991. https://doi.org/10.1016/j.iheduc.2024.100991

Relente, A. R. R., & Capistrano, E. P. S. (2024). Innovation self-efficacy, theory of planned behavior, and entrepreneurial intentions: The perspective of young Filipinos. Asia Pacific Management Review, 30 (3), 100350. https://doi.org/10.1016/j.apmrv.2024.100350

Renko, M., Shrader, R. C., & Simon, M. (2012). Perception of entrepreneurial opportunity: a general framework. Management Decision, 50(7), 1233-1251. https://doi.org/10.1108/00251741211246987

Romero-Colmenares, L. M., & Reyes-Rodríguez, J. F. (2022). Sustainable entrepreneurial intentions: Exploration of a model based on the theory of planned behaviour among university students in north-east Colombia. The International Journal of Management Education, 20(2), 100627. https://doi.org/10.1016/j.ijme.2022.100627

Roundy, P. T. (2022). Artificial intelligence and entrepreneurial ecosystems: Understanding the implications of algorithmic decision-making for startup communities. Journal of Ethics in Entrepreneurship and Technology, 2(1), 23–38. https://doi.org/10.1108/JEET-07-2022-0011

Schlaegel, C., & Koenig, M. (2014). Determinants of Entrepreneurial Intent: A Meta–Analytic Test and Integration of Competing Models. Entrepreneurship Theory and Practice, 38(2), 291-332. https://doi.org/10.1111/etap.12087

Shanock, L. R., Baran, B. E., Gentry, W. A., Pattison, S. C., & Heggestad, E. D. (2010). Polynomial regression with response surface analysis: A powerful approach for examining moderation and overcoming limitations of difference scores. Journal of Business and Psychology, 25(4), 543–554. https://doi.org/10.1007/s10869-010-9183-4

Shapero, A., & Sokol, L. (1982). The social dimensions of entrepreneurship. In C. Kent, D. Sexton, & K. V. Vesper (Eds.), The encyclopedia of entrepreneurship (pp. 72–90). Prentice Hall.

Sharma, L., Bulsara, H. P., Bagdi, H., & Trivedi, M. (2023). Exploring sustainable entrepreneurial intentions through the lens of theory of planned behaviour: A PLS-SEM approach. Journal of Advances in Management Research, 21(1), 20–43. https://doi.org/10.1108/jamr-01-2023-0006

Sharma, L., Bulsara, H. P., Trivedi, M., & Bagdi, H. (2023). An analysis of sustainability-driven entrepreneurial intentions among university students: the role of university support and SDG knowledge. Journal of Applied Research in Higher Education, 16(2), 281-301. https://doi.org/10.1108/jarhe-11-2022-0359

Srivastava, M., Shivani, S., & Dutta, S. (2024a). An empirical contribution towards measuring sustainability-oriented entrepreneurial intentions: A study of Indian youth. Environment, Development and Sustainability, 26(3), 7319–7345. https://doi.org/10.1007/s10668-023-03010-9

Srivastava, M., Shivani, S., & Dutta, S. (2024b). Sustainability-oriented entrepreneurial intentions: work values and the theory of planned behaviour. Journal of Small Business and Enterprise Development, 31 (2), 298-324. https://doi.org/10.1108/jsbed-03-2023-0105

Tan, L. P., Pham, L. X., & Bui, T. T. (2020). Personality Traits and Social Entrepreneurial Intention: The Mediating Effect of Perceived Desirability and Perceived Feasibility. The Journal of Entrepreneurship, 30(1), 56-80. https://doi.org/10.1177/0971355720974811

Tran, V. T., Nguyen, T. H., St-Jean, É., Duong, C. D., & Trinh, T. N. (2023). Social Entrepreneurial Intention among Youth in Vietnam: The Roles of Prior Experience and Perceived Educational Support. Journal of Social Entrepreneurship, 0(0), 1-29. https://doi.org/10.1080/19420676.2023.2244985

Truong, H. T., Le, T. P., Pham, H. T. T., Do, D. A., & Pham, T. T. (2022). A mixed approach to understanding sustainable entrepreneurial intention. The International Journal of Management Education, 20(3), 1-14. https://doi.org/10.1016/j.ijme.2022.100731

Ueda, D., Walston, S. L., Fujita, S., Fushimi, Y., Tsuboyama, T., Kamagata, K., Yamada, A., Yanagawa, M., Ito, R., Fujima, N., Kawamura, M., Nakaura, T., Matsui, Y., Tatsugami, F., Fujioka, T., Nozaki, T., Hirata, K., & Naganawa, S. (2024). Climate change and artificial intelligence in healthcare: Review and recommendations towards a sustainable future. Diagnostic and Interventional Imaging, 105(11), 453-459. https://doi.org/https://doi.org/10.1016/j.diii.2024.06.002

van der Westhuizen, T., & Goyayi, M. J. (2019). The influence of technology on entrepreneurial self-efficacy development for online business start-up in developing nations. The International Journal of Entrepreneurship and Innovation, 21(3), 168-177. https://doi.org/10.1177/1465750319889224

Vinuesa, R., Azizpour, H., Leite, I., Balaam, M., Dignum, V., Domisch, S., Felländer, A., Langhans, S. D., Tegmark, M., & Fuso Nerini, F. (2020). The role of artificial intelligence in achieving the sustainable development goals. Nature Communications, 11(1), 233. https://doi.org/10.1038/s41467-019-14108-y

Vuorio, A. M., Puumalainen, K., & Fellnhofer, K. (2017). Drivers of entrepreneurial intentions in sustainable entrepreneurship. International Journal of Entrepreneurial Behavior & Research, 24(2), 359-381. https://doi.org/10.1108/ijebr-03-2016-0097

Wang, Q., Li, Y., & Li, R. (2025). Integrating artificial intelligence in energy transition: A comprehensive review. Energy Strategy Reviews, 57, 101600. https://doi.org/10.1016/j.esr.2024.101600

Wang, Z., Deng, Y., Zhou, S., & Wu, Z. (2023). Achieving sustainable development goal 9: A study of enterprise resource optimization based on artificial intelligence algorithms. Resources Policy, 80, 103212. https://doi.org/10.1016/j.resourpol.2022.103212

Waris, I., Barkat, W., Ahmed, A., & Hameed, I. (2021). Fostering sustainable businesses: understanding sustainability-driven entrepreneurial intention among university students in Pakistan. Social Responsibility Journal, 18(8), 1409-1426. https://doi.org/10.1108/srj-10-2020-0399

Webometrics. (2023). Ranking Web of Vietnam Universities. https://www.webometrics.info/en/asia/vietnam

Yi, G. (2020). From green entrepreneurial intentions to green entrepreneurial behaviors: the role of university entrepreneurial support and external institutional support. International Entrepreneurship and Management Journal, 17(2), 963-979. https://doi.org/10.1007/s11365-020-00649-y

Zaremohzzabieh, Z., Ahrari, S., Krauss, S. E., Samah, A. A., Meng, L. K., & Ariffin, Z. (2019). Predicting social entrepreneurial intention: A meta-analytic path analysis based on the theory of planned behavior. Journal of Business Research, 96, 264-276. https://doi.org/10.1016/j.jbusres.2018.11.030

Zhao, H., Seibert, S. E., & Hills, G. E. (2005). The mediating role of self-efficacy in the development of entrepreneurial intentions. Journal of Applied Psychology, 90(6), 1265-1272. https://doi.org/10.1037/0021-9010.90.6.1265

Biographical notes

Son Tung Ha, Associate Professor and Dean of the Faculty of Business Administration at the National Economics University, focuses his research on entrepreneurship and sustainability-oriented behavior.

Thi Thanh Hoa Phan (Ph.D.), Researcher/Lecturer at the Faculty of Business Management, National Economics University, Vietnam. Her research interests include entrepreneurship and organizational behaviour.

Thi Viet Nga Ngo (Ph.D.), Researcher/Lecturer at the Faculty of Business Management, National Economics University, Vietnam. Her research interests include entrepreneurship and knowledge sharing.

Cong Doanh Duong (Ph.D.), Associate Professor at the Faculty of Business Management, Head of Lab for Business Analysis and Simulation, School of Business, National Economics University. His area of scientific interest includes entrepreneurship, corporate social responsibility, and sustainable development. His research on entrepreneurship has appeared in several journals, including the Journal of Retailing and Consumer Services, Personality and Individual Differences, Asia Pacific Journal of Marketing and Logistics, Education + Training, and Management Decision. He is an associate editor and a member of the editorial board of some journals, such as The International Journal of Management Education (Elsevier, SSCI/Q1, IF 5.2), Entrepreneurial Business and Economics Review (ESCI (ISI)/ Scopus Q1), Oeconomia Copernicana (SSCI IF = 8.5/Scopus Q1), and International Entrepreneurial Review.

Ngoc Thang Ha (Ph.D.), Researcher/Lecturer at the Faculty of Business Management, National Economics University, Vietnam. His research interests include entrepreneurship and online shopping.

Author contributions statement

Son Tung Ha: Conceptualization, Methodology, Investigation, Writing – Original Draft. Thi Thanh Hoa Phan: Conceptualization, Methodology, Formal Analysis, Writing – Review & Editing. Thi Viet Nga Ngo: Investigation, Writing – Review & Editing. Cong Doanh Duong: Conceptualization, Writing – Review & Editing, Data Collection, Writing – Original Draft. Ngoc Thang Ha: Investigation, Writing – Review & Editing.

Conflicts of interest

The authors declare no conflict of interest.

Citation (APA Style)

Ha, S.T., Phan, T.T.H., Ngo, T.V.N., Duong, C.D., & Ha, N.T. (2025). Integrating artificial intelligence competencies into the theory of planned behavior: Explaining sustainability-oriented entrepreneurial intentions. Journal of Entrepreneurship, Management and Innovation, 21(4), 30-53. https://doi.org/10.7341/20252142

---

Received 16 April 2025; Revised 10 July 2025; 13 September 2025; Accepted 17 September 2025.

This is an open-access paper under the CC BY license (https://creativecommons.org/licenses/by/4.0/legalcode).