The role of AI chatbots in fostering sustainable tourism destinations’ brand image and authenticity in Vietnam Open Access

Artificial intelligence (AI) is reshaping how travelers interact with destinations, particularly through conversational agents such as chatbots. These AI systems assist tourists in searching for information, personalizing recommendations and enhancing decision-making experiences before visiting a destination (Adamopoulou and Moussiades, 2020; Ukpabi et al., 2019). As the tourism sector accelerates its digital transformation, chatbots have become vital tools that connect tourists with destinations in real time, offering convenience and engagement across various touchpoints (Tosyali et al., 2023; Chi, 2023; Orden-Mejía et al., 2024). Sustainability has become an essential pillar in tourism development. For emerging economies such as Vietnam, where balancing tourism growth with environmental protection remains challenging, technology can play a transformative role (Pillai and Sivathanu, 2020; Santarém et al., 2018). Recent studies have acknowledged the potential of AI-based systems in promoting responsible consumption, destination transparency and traveler engagement with sustainable values (Liu et al., 2019; Khan et al., 2024a, b).

A significant theoretical and practical groundwork has been laid through investigations into the integration of chatbots in the tourism sector, but several gaps remain unexamined. First, according to Benaddi et al. (2024), chatbots currently have a substantial impact on travel services, functioning as a pivotal element in improving tourists' satisfaction levels as well as their overall experience quality, nonetheless, investigations into this area remain insufficient investigating the direct role of chatbot-provided information-related factors and tourists' interaction with AI chatbots on their perceptions and future behavior. Second, most recent studies on AI chatbots within the tourism sector, particularly in sustainable tourism, consider the role of chatbots in various aspects, including how chatbots influence tourists' continued usage intention (Jha et al., 2023; Pham et al., 2024), enhance their post-trip experiences (Majid et al., 2024a, b), inspire environmentally conscious behavior (Majid et al., 2024a, b), promote their willingness to pay for environmental conservation (Chi, 2024) or advocate for green destinations as part of sustainable tourism development (Arora and Chandel, 2024). However, the function of chatbots in promoting destination brand image and boosting destination brand authenticity to stimulate tourist visits and elevate their propensity to spend more to travel to this destination remains underexplored, although this constitutes a fertile ground for further scholarly exploration (Escobar-Farfán et al., 2024). Besides, prior studies have indicated that tourists' connection with nature contributes to encouraging sustainable tourism behavior and motivates tourists to safeguard the environment (Whitburn et al., 2019). Moreover, Jackson et al. (2013) highlighted that personal innovativeness is a key factor in adopting technological innovations, with user perceptions and subjective norms acting as intermediaries for behavioral intentions. However, in the context of AI-integrated tourism, how both personal connections with nature and personal innovativeness might influence the relationship between destinations' brand attributes and tourist behavior remains an unresolved question. Last but not least, to our knowledge, this study stands as one of the first initiatives to incorporate the SOR and ELM paradigms in order to explain how information-related factors provided by AI chatbots and tourists’ interactions with chatbots influence their perceptions and future behavior in the domain of tourism sustainability.

To address these gaps, this study investigates how information-related elements (quality, diagnosticity and relevance) and interaction-related elements (perceived enjoyment, accountability and anthropomorphism) of AI chatbots influence the brand image and authenticity of sustainable tourism destinations in Vietnam. It further examines how these brand attributes shape tourists' visit intention and willingness to pay a premium for sustainable experiences, as well as how personal innovativeness and nature connectedness moderate these effects. By integrating insights from the Stimulus–Organism–Response (SOR) and Elaboration Likelihood Model (ELM) frameworks, this study contributes a focused theoretical perspective on how AI chatbots can foster sustainable destination branding. The findings also provide actionable implications for tourism managers aiming to leverage AI technologies in support of the United Nations' Sustainable Development Goals (SDG 11 and 12).

One important dual-process model that explains how attitudes shift and develop is the elaboration likelihood model (ELM), which was first conceptualized and developed by Petty et al. (1981). In order to explain human attitudes, this ELM outlines the roles of core and peripheral routes. The core pathway, which stands for the cognitive method of information processing, shows how a person's attitude and decision-making have changed in response to the argumentative material presented. The peripheral route, on the other hand, refers to the external cues that affect a person's mindset and necessitate less mental work to absorb information. Compared to the processes involving emotive ties in the peripheral route, the central route typically involves a more deliberate examination of object-relevant arguments (Tang et al., 2012). Individuals pursue two separate paths when they are presented with persuasive stimuli: the central route and the periphery route. The primary cognitive processing path is crucial since recipients are required to carefully evaluate the message's contents, which improves the process of changing their attitude. As an alternative, the peripheral approach presumes the use of passive processes, in which individuals rely on factors like hierarchy, communicator gender, facial clues, among other factors (Zhang et al., 2024). Consequently, by taking into account the skills and motivations of the intended audience, this model aids in creating a compelling message. SanJosé-Cabezudo et al. (2009). It has been proven that external information plays a significant part in shaping customer attitudes and behavior modification through the ELM theory. In a previous article, they used ELM to assess the reliability, accuracy of recommendations from AI chatbots with the aim of assessing their effect on consumers' consciousness and desire to adopt new technology in Zhang et al. (2024) or depending on whether or not they receive adequate information, a consumer's attitude may shift from positive to negative or strong to weak (Nolder and Kadous, 2018). From the above evidence, we can confirm that ELM has become an influential theory and has contributed highly to understanding the phenomenon of persuasion in many different fields, including political science, communication science, marketing communications, etc. As it is evident that the ELM theory has been extensively used in earlier studies to address AI chatbots in the area of tourism, our study will yield multiple benefits from those investigations, keeping up the application of the ELM theory to illustrate customer emotions and interactions with AI chatbots in the tourism sector.

This article applies ELM theory to explain AI chatbot recommendation information – related factors, tourist – AI chatbot interaction factors, nature connected, and personal innovation that affect customer perceptions of sustainable destination brand authenticity (SDBA) and image, which in turn affects awareness and future behavior. Specifically, visitors of interactions with AI chatbots and Nature Connected (NC) and Personal Innovation (PI) are regarded as peripheral routes, whereas AI chatbot information-related elements are regarded as the main route.

First proposed by Mehrabian and Russell (1974) and subsequently enhanced by Jacoby (2002), the stimulus-organism-response (SOR) theory provides the theoretical basis for this research. According to Mehrabian and Russell (1974), stimuli (S) present in particular environments shape individuals' perceptions (O), ultimately resulting in specific responses (R). Additionally, Jacoby (2002) emphasized that personal behavior is influenced by an individual's internal state. Expanding on this framework, Kim and Lennon (2013) incorporated both external stimuli, such as information from external sources, and internal stimuli, such as website quality, demonstrating their impact on consumer responses. Numerous travel advertisements, videos, and television programs are structured to guide the thought process and feelings of the audience, thereby stimulating desires and encouraging them to make decisions to visit the destinations being featured (Jiang et al., 2016).

The S-O-R model has been employed to explore how technologies like virtual reality and chatbots affect tourists' emotions and predict their behavior within the framework of sustainable tourism (Pham et al., 2024; Kim et al., 2020). Research in this domain remains relatively scarce. Therefore, based on previous studies, this research considers AI-related factors (including quality, quantity, diagnosticity, relevance and value-added) as stimulus factors that influence perceptions of authenticity and destination brand image (Baloglu and McCleary, 1999; Manhas et al., 2016). Sustainable destination brand image (SDBI) and SDBA are central to the proposed model and are considered the organismic factors. Finally, visit intention and willingness to pay a premium are regarded as the response components of this study framework.

Destination brand image is a set of characteristic elements of a destination that is constituted of physical characteristics, sociocultural factors and promotional communication (Walmsley and Jenkins, 1993; Echtner and Ritchie, 1993; Embacher and Buttle, 1989; Hankinson, 2005). It contributes to distinguishing the destination while playing a pivotal role in attracting tourists (Kladou and Kehagias, 2014; Ruiz-Real et al., 2020). With current trends, the connection between destination brand image and “sustainable” is increasingly being attended to when tourists often choose destinations that they feel these destinations make a positive contribution to the economy, social culture, environment, community and they are created the opportunity to directly perform responsible behavior toward the environment (Lee et al., 2013; Lee and Jan, 2019). Combining the above factors, in this research article, SDBI is defined as the brand image of a destination being remembered in a positive way according to the public by the difference compared to other destinations (Qu et al., 2011); simultaneously, it meets the current needs of tourists, and create value for local communities through sustainable development values in terms of environment, economy, culture and society (Curtin and Busby, 1999).

Within the tourism field, destination brand authenticity is a multidimensional concept. As stated by Manthiou et al. (2018), authenticity serves a key function in enhancing tourists' impression and brand love toward the destination. Destination brand authenticity is not only the objective factors, such as history and culture that motivate tourists to destination, but also includes their subjective experiences (MacCannell, 1973; Wang, 1999). The way to know whether the destination brand image is built in accordance with reality or not and whether it meets or exceeds tourists' expectation must be based on their perceptions and experiences (Zhang and Yin, 2020). Based on Ramkissoon and Uysal (2011) findings, a destination is considered authentic when it retains its characteristic elements without being “over-commercialized”, thereby building trust and attracting tourists eager to explore the destination. To ensure sustainability, destination brands not only preserve but also develop their core value, while satisfying tourists' demands in a harmonious way (Chen et al., 2020).

In our research, the factors related to AI chatbot-recommended sustainable tourism information are characterized by five dimensions: quality, quantity, diagnosticity, relevance and value-added. First, the definition of information quality is the caliber of information that a system generates (DeLone and McLean, 1992). Our study refers to this factor as the quality of information and recommendations that AI chatbots can offer to users and is characterized by four dimensions: accuracy, comprehension, timeliness and currency (Thomas et al., 2019). Some earlier literature demonstrates that the quality of information, has a crucial position in shaping customers' trust (Zhao et al., 2020; Xuan Cu Le, 2023) and their perceptions of a product, service, or even a brand in general (Rodrígueza et al., 2019; Li et al., 2023). In accordance with Morhart et al. (2015) and Safeer et al. (2022), in tourism, when an AI chatbot delivers accurate, detailed, timely and up-to-date information about recommended sustainable destinations, it can shape positive and trustworthy images in the consumer's mind, which contributes brand authenticity. Moreover, if this information truly reflects the destination's essence and character, it enhances tourists' brand authentic experiences because they suppose this destination brand is consistent (Schallehn et al., 2014). For these reason, we hypothesize that:

Second, under the circumstances of shoppers using online feedback, the quantity of information refers to the amount of online reviews about a service or product (Filieri, 2015). In the same way, within our research, the information quantity of AI chatbots is considered to be the volume of information and recommendations about sustainable destinations that these chatbots can furnish to users. As indicated by Lopes et al. (2020), Purnawirawan et al. (2014) and Thomas et al. (2019), the quantity of information significantly affects consumers' perception of products' usefulness and credibility in the context of online reviews. In addition, Baloglu and McCleary (1999) have demonstrated through empirical evidence that an adequate amount of information positively influences destination image formation, which is the foundation of destination brand image (Manhas et al., 2016). Thus, we hypothesize that:

Third, information diagnosticity is described as a website's ability to enhance users' perception or knowledge about particular services or products (Andrews, 2013), thereby assisting them in evaluating these items (Jiang and Benbasat, 2004). In our study, this term is used to describe an AI chatbot's capability to recommend sustainable destinations and provide their information, which can help users assess the quality and performance of the recommended destinations. Literally, the diagnosticity attribute helps users to easily and quickly assess offerings’ performance and get acquainted with them (Henni et al., 2022; Filieri, 2015; Jiang and Benbasat, 2004). Most AI chatbot users seek immediate and consistent information (Brandtzaeg and Følstad, 2017). Same as with sustainable tourism AI chatbots, the high level of diagnosticity of provided information can establish an initial image of destination brand image and authenticity, especially about brand's performance and brand promise. Accordingly, we hypothesize that:

Fourth, information relevance represents the level at which information is appropriate and helpful in accomplishing a specific task (Filieri and McLeay, 2014). In the context of our study on AI chatbots about sustainable destinations, we considered this term as the degree to which the provided recommendations and information are relevant to users. Mishra et al. (1993) found that information relevance performs as a key player that impacts attraction effect, which describes the increased likelihood that consumers will choose the target brand when an asymmetrically inferior alternative is introduced (Huber et al., 1982). This can be adapted in our research that if an AI chatbot can provide relevant information and recommendations, which are suitable for tourists' demand, it will increase their positive impression about the destination. Besides, previous studies have pointed out that information relevance and value-added, as a component of information quality in online reviews context, bring positive results on the recommended destination's image (Rodrígueza et al., 2019; Kim et al., 2017). On the other hand, when AI chatbots can provide a recommendation that closely meets tourist needs and expectations, it will reinforce the impression that the destination brand is both genuine and thoughtfully represented. Thus, we hypothesize that:

Last but not least, information value-added is understood as the level of usefulness and advantages that information brings to users (Wang and Strong, 1996). In our research, this factor describes the extent to which information and recommendations about sustainable destinations provided by AI chatbots are beneficial and instrumental to tourists. As mentioned, Kim et al. (2017) and Rodríguez et al. (2019) findings demonstrate that value-added information considerably impacts the destination image. In addition, AI chatbots can create more additional information by gaining user's insight from analyzing historical users' data processes (Akhtar et al., 2019). These additional insights can be beneficial for tourists to raise knowledge about both the positive and negative sides, well-known and unknown things of destinations, and consolidate tourists' perception of the recommended sustainable destinations brand. Thus, we hypothesize that:

Firstly, perceived enjoyment is recognized as an intrinsic motivator that emphasizes the process of use and reflects the joy and pleasure derived from using a system (Merikivi et al., 2016). Previous studies have highlighted how brands and businesses now use chatbots as tools to provide 24/7 customer support. Findings indicate that the enjoyment experienced after interacting with technology-driven information significantly affects customer satisfaction, which, in turn, positively influences their goodwill toward the brand image and business (Ashfaq et al., 2020; Selamat and Windasari, 2021; Ding et al., 2022). Within the realm of tourism, a fun, humorous, and engaging guide contributes to creating a lasting impression of the destination image in the tourists' perceptions (Min, 2012; Li et al., 2022). Likewise, in our study, perceived enjoyment refers to the extent of fun, interest, and satisfaction that users experience after engaging with and receiving destination-related information provided by an AI chatbot.

Research on the online authenticity of virtual attractions and destinations indicates that technology users do not need objective proof in order to experience authenticity. Despite being influenced by technology, the elements at a destination can still convey sufficient meaning, leading users to perceive it as authentic (Hall, 2007; Jiménez-Barreto et al., 2020). Rouibah et al. (2016) assert that enjoyment is likely to have a beneficial impact on is anticipated to have a positive effect on trust enjoyment has a positive impact on trust, serving as two key drivers that promote the adoption of technology and help mitigate users' perceived risks in utilizing technology. Moreover, customers have trust in a brand; it not only strengthens their bond with the brand but also positively impacts their perception of its sincerity and authenticity (Kushwaha et al., 2021; Taheri et al., 2019). Similarly, the information provided by the AI chatbot may be accurate or not entirely precise, but during the interaction, users perceive humor and interest, leading to their satisfaction with the chatbot, which in turn is likely to gradually strengthen users' trust and enhance the brand's authenticity. Therefore, we hypothesized that:

Second, perceived algorithmic accountability requires an explanation of the user's beliefs, actions or emotions related to the algorithm, and users expect organizations to implement an evaluation process with positive or negative consequences depending on the task outcome (Shin and Park, 2019). In our research, perceived accountability refers to the extent to which travelers perceive that the organization providing the chatbot will be held responsible and take measures to support users when they encounter issues with the information or suggestions related to sustainable destinations provided by the AI chatbot during its use. If a brand is accountable for the destination information it provides, it will not only increase the reliability of the travel details but also showcase the brand's commitment. This, in turn, helps create a positive perception of the brand's image (Mohammed and Rashid, 2018; Hassan and Soliman, 2020; Aysolmaz et al., 2023). The explosion of information on the Internet, combined with excessive commercialization and deceptive advertising strategies, has led to a sense of meaninglessness in society (Morhart et al., 2015). To alleviate this uncertainty, many individuals seek authenticity as a way to restore trust (Li et al., 2024). Research on information suggested by AI chatbots indicates that when the information is perceived as being highly accurate and reliable, the users' trust in the service is reinforced, thereby strengthening their travel decisions (Kim et al., 2023; Ali et al., 2023), which also contributes to enhancing the brand's authenticity (Kim et al., 2020). Based on these points, we theorize that:

Lastly, anthropomorphism involves ascribing human traits such as emotion, behavior, personality, etc. To non-human instances like animals, things, or even natural phenomena (Epley et al., 2007). In our research, anthropomorphism can be understood as the extent to which tourists recognize an AI chatbot as capable of interacting like an actual human while they seek sustainable destination information and recommendations via the chatbot. Previous research highlights that anthropomorphism is an indispensable component of brand attachment, brand experience and building intimate relationships with customers (Chen and Lin, 2021). Specifically, anthropomorphic chatbots can enhance customers' experiences (Rizomyliotis et al., 2022; Kayeser Fatima et al., 2024) and increase their emotional connections with brands (Araujo, 2018) in various fields such as tourism, e-commerce, etc. Users' perceptions of chatbot anthropomorphism are shown to significantly influence users' trust in chatbot (Konya-Baumbach et al., 2023; Khan et al., 2024a, b), which in turn impacts the brand trustworthiness, an important element of brand authenticity (Safeer et al., 2022). However, a survey conducted by Yuan et al. (2022) with respondents who have interacted with Xiaomi's AI-driven customer service assistant in China found that the agent's human-like characteristics have no significant impact on brand image despite playing a vital role in brand promotion. Hence, in the scope of promoting sustainable tourism destination brands, we hypothesize that:

In the scope of tourism, numerous investigations have confirmed the greater willingness to pay a premium for attractions that are perceived as sustainable. For example, Reynisdottir et al. (2008) confirmed this trend in natural attractions, while Hinnen et al. (2015) highlighted its significance in ecotourism, and López-Sánchez and Pulido-Fernández (2016) emphasized the critical role of sustainable destinations. Line and Hanks (2015) further demonstrated that tourists are open to paying higher prices for destinations perceived to invest in environmental protection and contribute positively to local communities. More recently, Li et al. (2023), in an empirical investigation conducted via the Credamo platform, emphasized that through communication with chatbot, brand image significantly impacts customer willingness to pay within the hospitality industry. Specifically, when users perceive the information provided by chatbot as useful, valuable, and enjoyable during the interaction, then it positively impacts on brand image in customers' perception, thereby fostering their more favorable future behavior (Yuan et al., 2022; Cheng and Jiang, 2021; Iranmanesh et al., 2024). Similarly, in this research endeavor, destination brand image can influence tourists' behavior, especially their readiness to pay a premium cost to travel to a sustainable destination that AI chatbot has recommended.

A well-crafted brand image also creates a strong impression on visitors and enhances their intention to visit (VI) (Baker and Cameron, 2008). The strong connection between destination brand image and travelers' VI was similarly confirmed by Kumail et al. (2021). In the current era of technology, Cheung et al. (2020) emphasized that destination brand communities on social media significantly shape tourists' emotions toward the brand, evoking joy, positive surprise and love, which later exert a substantial influence on their inclination to visit. Moreover, Lou et al. (2024) asserted that AI not only enhances cognitive and emotional connections with destinations but also supports travelers in making decisions to choose sustainable destinations for tourism activities. Therefore, we hypothesize that:

With the growing influence of AI technology in tourism marketing (Benaddi et al., 2024), sophisticated informational tools have contributed significantly to reinforcing the authenticity of commercial signals. AI technologies, including chatbots and personalized recommendation systems, are designed to strengthen perceptions of authenticity by delivering tailored interactions (Ku, 2024). Moreover, Ghorbanzadeh et al. (2025) highlight that effectively integrating social presence through AI fosters immersive and engaging experiences, which in turn enhances the positive connection between tourists' experiences and the destination's brand.

Brand authenticity is a crucial element in enhancing the value of a destination's brand and increasing customers' willingness to pay premium prices (Fatma and Khan, 2024). It has been demonstrated in several studies that tourists are willing to spend more on experiences that are personalized and unique, reflecting local cultural values (Hennig-Thurau, 2004). Clear perceptions of brand authenticity can increase consumers' willingness to pay by aligning with their values and preferences for sustainability (Kim and Huang, 2021). Additionally, studies by Morhart et al. (2015), Manthiou et al. (2014), and Eggers et al. (2013) confirm a positive link between brand authenticity and tourist engagement. Specifically, when the authenticity of a destination brand is perceived, tourists generally show more positive attitudes and a greater desire to explore the destination (Schallehn et al., 2014). This indicates that authenticity is not merely a branding objective but also a critical factor in fostering long-term relationships with tourists and influencing their VI (Kumail et al., 2021). Therefore, we hypothesize that:

Strong connectedness to nature allows visitors to reveal more love and respect for the environment, which greatly influences their travel choices and actions (Whitburn et al., 2019; Shimul et al., 2024). According to Mayer and Frantz (2004), a person's sense of nature connectedness reflects a mental and experiential link to the surroundings. In particular, when interacting with AI chatbots in their travel information search, these tourists tend to appreciate and trust more information related to environmental protection and sustainable tourism practices. A SDBI conveys a destination's commitment to environmental conservation and sustainable tourism practices (Hanna et al., 2018; Sartori et al., 2012). For tourists with a strong connection to nature, this imagery, especially when delivered through highly interactive and personalized AI chatbots, is likely to resonate more deeply with their environmental values and personal identity (Nisbet et al., 2009).

As stated by Kelly (2018) and Pritchard et al. (2020), people who have an intense attachment to nature are more likely to support environmental initiatives, even if they require greater financial outlays or personal sacrifices. This may be further reinforced when they receive quality and trustworthy information from AI chatbots about the destination's sustainability initiatives. Van der Linden (2015) and Tam (2013) found that individuals who feel connected to nature not only demonstrate higher trust in sustainable practices but are also more willing to invest in environmental conservation and exhibit more environmentally friendly behaviors. This factor can be enhanced by the AI ​​chatbot providing highly diagnostic and value-added information about the destination's environmental aspects. This is further confirmed by research by Shimul et al. (2024), which found that tourists who have a strong connection to nature, when supported by AI chatbots capable of expressing empathy and responsibility, are more likely to choose and support environmentally responsible travel options.

Additionally, tourists with strong nature connections develop heightened environmental identities that influence consumption patterns and travel behaviors (Perkins and Brown, 2012). This connection results in deeper emotional responses to brand images associated with sustainable practices, significantly impacting their visit intentions (Spence et al., 2018). When sustainable brand images evoke positive affective responses, such as attachment, individuals with strong nature connectedness show an increased desire to visit these destinations (Tussyadiah et al., 2018). As a result, we hypothesize that:

The strong bond with nature enhances tourists' ability to discern and value authentic environmental initiatives (Dwyer et al., 2009). Authentic environmental practices, when perceived as genuine, make individuals with nature-connectedness more receptive to premium pricing (Kadirov, 2015). As mentioned previously, by providing customized interactions, artificial intelligence tools such as chatbots and tailored recommendation engines are made to increase people's sense of authenticity (Ku, 2024). In addition, Line and Hanks (2015) showed that tourists are willing to pay an extra fee that is thought to make investments in environmental preservation and benefit local communities. As a result, for nature-connected individuals, brand authenticity will be a critical determinant in evaluating premium pricing for sustainable destinations through AI chatbot recommendations when they perceive the sustainability of this destination. Furthermore, in tourism, authentic brand practices can trigger both cognitive and affective responses, influencing visit intentions. Kumail et al. (2021) found that tourists with strong nature connectedness, responding to genuine environmental efforts, are more likely to experience positive emotional responses, further shaping their VI. Consequently, we hypothesize that:

According to Turan et al. (2015) and Ayanwale and Ndlovu (2024), personal innovativeness is the willingness of an individual to try new technologies, such as AI chatbots, and concepts when they feel that they are consistent with their values and beliefs. Moreover, customer attitudes and behavior are significantly impacted by personal innovativeness (Lu et al., 2005). AI chatbots, through their ability to provide transparent information, real-time communication and personalized interactions, enhance travel experience (Reddy et al., 2024; Doğan and Niyet, 2024). However, individuals may not all respond to SDBI in the same way, with personal innovativeness acting as an important moderator. People with higher levels of personal creativity may perceive SDBI as more appealing, leading to being open to spending more on unique sustainable experiences.

Furthermore, highly creative people are more inclined to gladly interact via AI chatbots to gain knowledge about the destination and explore the unique features and sustainable values of the destination through conversations (Tosun et al., 2024). Research by Zhu et al. (2023) and Reddy et al. (2024) shows that users are better able to understand the destination's features when they exchange detailed information with AI chatbots, which increases the likelihood that they will visit. For this reason, the more the destination brand image reflects sustainability, the stronger the intention of highly creative people to choose to visit that destination.

As mentioned before, tourists typically have more positive views and a stronger willingness to explore a destination when they believe the brand to be authentic (Schallehn et al., 2014). This raises both the intention to travel and the readiness to pay more cost to obtain sustainable destinations, especially among tourists with high levels of personal innovativeness. Therefore, we hypothesize:

All measurement items were adapted from validated scales and translated into Vietnamese using a back-translation process to ensure semantic accuracy. A pilot test with twenty respondents was conducted to evaluate clarity and content validity, leading to minor wording adjustments without altering the conceptual meaning of the items.

The target population included individuals who had previously used AI chatbots in a tourism context. Because this group is difficult to identify in the general population, an online survey was distributed via Google Forms using snowball sampling, a common nonprobability method that enables efficient access to specific users. To enhance diversity within the sample and reduce the risk of homogeneity, the survey was distributed across multiple online platforms (e.g. social media groups, travel forums and university networks). To ensure that only relevant participants proceeded to the main questionnaire, two screening questions were included: (1) “Have you ever used a chatbot before?” and (2) “Have you ever used a chatbot for travel advisory to sustainable tourism destinations?”. Only respondents who answered “Yes” to both questions were allowed to continue. This ensured that only individuals with relevant experience were included in the final sample by confirming their prior chatbot use in a tourism context. The full survey will take approximately 10 min to complete. In addition, the survey introduction explicitly clarified that the study focused on experiences with AI chatbots providing sustainable or eco-friendly destination recommendations. This clarification ensured that only individuals with relevant and authentic chatbot experience related to destination sustainability participated in the main survey. It is important to clarify that respondents were not asked to evaluate the objective sustainability performance of any destination, country, or region. Instead, the survey explicitly focused on respondents' perceptions of sustainability-related attributes as communicated through AI chatbot recommendations during the pre-visit stage. In other words, participants evaluated sustainability cues embedded in AI-mediated destination information, such as environmental responsibility, ethical positioning and community-oriented values, rather than making judgments about destinations' actual sustainability outcomes. This perception-based framing is consistent with prior destination branding and digital tourism studies, which conceptualize sustainability as a cognitively constructed image shaped by mediated information exposure rather than as an objective assessment criterion.

A total of 783 responses were collected between October 7th and October 21st, 2024. After removing incomplete or extreme responses, 508 valid cases (Table 1) were used to evaluate the suggested model (Figure 1), with an effectiveness rate of 64.9%. The output is quite balanced in terms of gender, including 214 men (42.1%) and 294 women (57.9%). As for the sample characteristics, the participants' ages are mostly from under 25 to 45, specifically, the most common subjects were under 25 years old, at 59.3%. Regarding the revenue source, 56.5% of respondents have monthly incomes below 10 million VND, 22.4% of them are from 10 to 20 million VND per month, 13.6% of them are from 21 million VND up to 30 million VND and 7.5% have an income over 30 million. In addition, 254 surveyors (50%) are residing in Ho Chi Minh City and working there, 113 surveyors (22.2%) are in Hanoi City, 62 surveyors (12.2%) are in Da Nang City, and the remaining come from other provinces and cities. This distribution reflects a mix of urban and regional perspectives and provides insight into different tourism contexts within Vietnam.

To test the theoretical model (Figure 1), scales were adopted from the literature on AI chatbot recommended sustainable tourism destination information-related factors which includes: information quality, information quantity, information diagnosticity, information relevance and information value-added; tourist AI chatbot interaction includes: perceived enjoyment, perceived accountability, perceived anthropomorphism, SDBI, sustainable destination authenticity, willingness to pay a premium price (WTPP) for sustainable destination, visit intention, nature connectedness and personal innovativeness.

All items included in the scale were translated from English to Vietnamese, aiming to ensure that the survey respondents grasped the information and their meanings correctly (Harkness et al., 2004). Respondents used a 7-point Likert scale to express their views for each question, ranging from: strongly disagree, disagree, somewhat disagree, neither disagree nor agree, somewhat agree, agree and strongly agree.

Information quality was modified from Thomas et al. (2019) (e.g. “Information via chatbots is comprehensive”). Information quantity was adapted from Zhang et al. (2014) and Filieri et al. (2018) (e.g. “There is a variety of product information via chatbots”). Information diagnosticity was adapted from Jiang and Benbasat (2004) (e.g. “The information provided by AI chatbot was helpful for me to evaluate the destination”). Information relevance was adapted from Mishra et al. (1993) (e.g. “I believe that the information and suggestions about sustainable tourism destinations provided by AI chatbot are useful to me”). Information value-added was adapted from Filieri and McLeay (2014) (e.g. “The information I obtain from AI chatbot enables me to detect unknown aspects of a specific sustainable tourism destination”). Perceived enjoyment was adapted from Ghani and Despande (1994) (e.g. “I found asking the AI chatbot for information and suggestions about sustainable tourism destinations interesting”). Perceived accountability was adapted from Wieringa (2020) and Kacianka and Pretschner (2021) (e.g. “I think that if this AI chatbot causes harm and nuisance for me, then this would be compensated by the organization providing the AI chatbot”). Perceived anthropomorphism was adapted from Bartneck et al. (2009) and Balakrishnan and Dwivedi (2021) (e.g. “Chatbot are natural; I do not feel fake about it”). SDBI was adapted from Lee and Lockshin (2011) (e.g. “I had a positive image of this sustainable tourism destination after experiencing the chatbot application”). Sustainable destination authenticity was adapted from Schallehn et al. (2014) (e.g. “Considering its brand promise, the sustainable tourism destination does not pretend to be something else”). WTPP for a sustainable destination was adapted from Sahin et al. (2011) and Casidy and Wymer (2016) (e.g. “I am willing to pay a lot more to visit this sustainable tourism destination than visit other destinations”). Visit intention was adapted from Huang et al. (2013), Jang and Namkung (2009), and Tussyadiah et al. (2018) (e.g. “I am willing to visit the sustainable tourism destination that I was recommended by AI chatbot soon”). Nature connectedness was adapted from Shimul et al. (2024) (e.g. “I often feel connected to nature”). Personal innovativeness was adapted from Jackson et al. (2013) (e.g. “I like to experiment with new information technologies”). All sustainability-related constructs in this study were operationalized as perception-based measures (Appendix). Specifically, SDBI and SDBA capture respondents' perceived sustainability-oriented attributes of destinations as inferred from AI chatbot interactions, rather than objective or verified sustainability indicators. This approach aligns with established tourism branding literature, which emphasizes that destination image and authenticity are formed through tourists' subjective interpretations of communicated cues and symbolic meanings, particularly in digitally mediated environments.

Following Lim et al. (2023), the partial least squares path modeling (PLSPM) method minimizes biased estimates and reduces potential errors, making it effective for handling unknown data. Therefore, this study applied it for comprehensive research frameworks that incorporate advanced constructs and moderators. To evaluate model fit, the bootstrapping technique was put in, as it thoroughly fixes multicollinearity concerns (Burlea-Schiopoiu et al., 2021). Additionally, the SEM-PLS approach was chosen based on its versatility in managing multiple independent variables concurrently, how well it works with primary and secondary datasets, and its suitability for tiny sample sizes. The study used PLS-SEM via SmartPLS 4.0 to test relationships, focusing on prediction rather than theory validation (Elshaer et al., 2024).

Common data collection methods can cause covariance between variables and will cause problems for the next steps in data analysis MacKenzie and Podsakoff (2012). This issue applies to our study as well, given that the data were collected at a single point in time, potentially introducing common method bias (CMB). To address this concern, Harman's single-factor test was employed to assess the presence of CMB (MacKenzie and Podsakoff, 2012). The results indicate that a single factor accounts for a maximum of 44.58% of the total variance, which is below the threshold of 50%. This suggests that the variance is not excessively influenced by a single factor. Therefore, it can be concluded that this study is not significantly affected by CMB, guaranteeing the objectivity and consistency of the findings.

Table 2 represents our evaluation of reliable values (outer loadings) result, which must exceed 0.708 (Hair et al., 2021). Obviously, all variable indicator loading has surpassed this threshold. Then, we examined the reliability and validity of latent variables using composite reliability (CR), Cronbach's alpha (CA), which must be above 0.7 according to Dijkstra and Henseler (2015), and the average variance extracted (AVE), which must be greater than 0.5 according to Rasoolimanesh et al. (2016). We find that all indicators are satisfactory. Therefore, with this result, we confirm that all variables' indicators exhibit positive correlations.

In particular, we find that CA and CR values of all variables are acceptable. Moreover, these acceptable CR and CA values indicate that these observed variables are stable and consistent in measuring the concept. In addition, all variables' AVE values represented in Table 2 have well surpassed the recommended threshold of 0.5; the lowest AVE value of the PAN variable at 0.675 is greater than the required standard value. These results demonstrate that our indicators effectively measure the latent variables, thus substantiating the reliability and validity of our model. All values were below the threshold of 0.90 for the heterotrait-monotrait ratio (HTMT) (Table 3), indicating good discriminant validity among constructs.

The square roots of AVE for each construct being greater than its correlations with other constructs further confirmed the discriminant validity, as derived from the Fornell–Larcker criterion (Table 4).

In Table 5, the analysis of latent variable correlations revealed moderate to strong positive relationships among constructs, supporting the theoretical framework of the study.

The path coefficients were evaluated using a nonparametric bootstrap technique with 5,000 resamplings in order to assess the precision of the structural model (Lim et al., 2023), based on the results presented in Table 6.

As findings outlined in Table 6 and Figure 2, among information-related factors, ID (β = 0.140, t = 2.382, p < 0.05) and IR (β = 0.181, t = 2.613, p < 0.01) significantly influenced SDBA, while both IQL (β = 0.134, t = 3.025, p < 0.01) and IR (β = 0.180, t = 2.993, p < 0.01) demonstrated positive effective effects on SDBI. PAN exhibited strong significant effects on both SDBA (β = 0.270, t = 5.152, p < 0.001) and SDBI (β = 0.179, t = 4.022, p < 0.001), while PE showed strong positive associations with SDBA (β = 0.194, t = 2.960, p < 0.01) and PAC showed a significant effect on SDBA (β = 0.099, t = 2.236, p < 0.05). The analysis of rejected paths from the hypothesis testing reveals several important patterns in the relationships between variables. IQL showed a negligible negative relationship with SDBA (β = −0.009, t = 0.184, p = 0.854), indicating virtually no direct connection between information quality and SDBA. Similarly, IVA demonstrated no significant influence on either brand authenticity or image, with both relationships being rejected (SDBA: β = 0.002, t = 0.030, p = 0.976; SDBI: β = 0.064, t = 1.304, p = 0.192). The quantity of information (IQT) also proved to be insignificant in impacting brand outcomes, with both paths to SDBA and SDBI being rejected (β = 0.041 and −0.031, respectively). An interesting pattern emerged with PAC and PE, where both variables significantly influenced SDBA but failed to establish significant relationships with SDBI (PAC → SDBI: β = 0.047, t = 1.456, p = 0.146; PE → SDBI: β = 0.063, t = 1.313, p = 0.189). These findings suggest that information characteristics generally have weak direct effects, while emotional and perceptual factors tend to have stronger impacts on brand authenticity than brand image. The rejection patterns indicate that SDBA might be more sensitive to various influencing factors compared to SDBI, with several variables showing asymmetric effects by influencing SDBA but not SDBI.

The explanatory power of the model was assessed through R-square values (Table 7). The model showed substantial explanatory power, accounting for 67.1% of the variance in SDBI (R² = 0.671), 61.6% in VI (R² = 0.616), 52.3% in SDBA (R² = 0.523) and 47.9% in WTPP (R² = 0.479).

The analysis (Table 6) revealed substantial relationships among the major constructs. SDBA exhibited strong positive effects on VI (β = 0.239, t = 3.910, p < 0.001) and WTPP (β = 0.388, t = 6.286, p < 0.001). Finally, SDBI demonstrated a significant positive contribution to VI (β = 0.285, t = 4.851, p < 0.001).

The model (Figure 1) includes interaction variables to investigate how two factors moderate effects. First, nature connectedness moderates how sustainable destination authenticity impacts SDBI concerning both visit intention and WTPP for a sustainable destination. Second, personal innovativeness moderates how sustainable destination authenticity influences SDBI regarding visit intention and WTPP for a sustainable destination.

The moderation results show that personal innovativeness affects the connections between SDBI and WTPP for a sustainable destination (PI × SDBI → WTPP, H12d: β = 0.172. p < 0.05). This shows that higher personal innovativeness strengthens the connection between an SDBI and users' WTPP for a sustainable destination. However, there is no moderating impact of personal innovativeness on the connection of SDBI and visit intention (PI × SDBI → VI, H12c: β = 0.019, p > 0.05). Regarding nature connectedness, the outcomes do not demonstrate the moderation nor the relation of SDBI and WTPP for a sustainable destination (NC × SDBI → WTPP, H11d: β = 0.041, p > 0.05) nor the connection of SDBI and visit intention (NC × SDBI → VI, H11c: β = −0.101, p > 0.05). Moreover, nature connectedness and personal innovativeness do not affect the relation between sustainable destination authenticity and WTPP and the connection between sustainable destination authenticity and visit intention. Consequently, only H12d was validated (see Table 8).

Our research introduces an empirical model (Figure 1) to investigate the effect of five AI-powered chatbots information-related factors (quality, quantity, diagnosticity, relevance and value-added) and AI chatbots-tourists interactions factors (perceived enjoyment, anthropomorphism and accountability) on SDBI and authenticity, then influence the tourists' visit intention and openness to pay more to travel there. In addition, we also examine the moderate role of natural connectedness and personal innovation on the relationship between SDBI, SDBA and VI, WTPP.

Our results indicate that three out of five AI chatbot information-related factors significantly influence SDBI and authenticity. Particularly, information quality and relevance impact on tourist's perception of SBDI, while information diagnosticity and relevance affect SDBA. Our findings partly match with Kim et al. (2017)'s findings, which demonstrate that relevance affects the destination's image, whereas quantity and value-added do not. In contrast, Rodrígueza et al., (2019) results showed that only information value-added has an effect on destination image, while relevance and quantity do not. The difference may come from our distinctive research context. The two mentioned papers are studied about the influence of information quality in the tourism sector from social media online reviews, such as Facebook pages in Kazakhstan and Sina Weibo in China. Literally, they investigated information relevance and value-added as two components that belong to the information quality criterion. Moreover, based on current knowledge, the impact of AI chatbots on the tourism sector is still being investigated by a limited number of researchers (Benaddi et al., 2024), especially the direct relationship between AI chatbot information elements and brand image as well as authenticity, particularly in relation to sustainable tourism.

This research discusses AI chatbot interactive elements that have a strong effect on SDBI. AI Chatbot is produced and designed to entertain customers with their export topics and tips, which has an important impact on the truly sustainable brand image of that destination. If customers have positive emotions, they will want to experience the place that the AI Chatbot has noticed and be willing to pay a premium to visit that certain destination over others. People are more likely to believe in AI if they trust its recommendations are accurate as well as trustworthy Zhang et al., (2024), Moradi and Ahmadian (2015). These results have significant ramifications for developing and deploying AI chatbots that guarantee perceived enjoyment, accountability and anthropomorphism user trust through the process of using AI Chatbot. As far as we are aware, there is little research examining the relationship between AI chatbot interactive elements and SDBI, especially within the framework of sustainable tourism. This will be our first research in this field.

Our research shows that both SBDA and SBDI demonstrated a significant contribution to WTPP and VI. When a destination is truly sustainable, its brand image is built in accordance with its commitment to its nature, not simply embellishment or “greenwashing”, so that tourists are inclined to pay higher prices to experience this sustainable destination hub, showcasing its appeal over alternative locations. This outcome aligns perfectly with the results found in studies by López-Sánchez and Pulido-Fernández (2016), Kiatkawsin and Han (2019), Fatma and Khan (2024). Moreover, as noted by Schallehn et al. (2014) and Kumail et al. (2021), when a destination brand image is perceived as remarkably authentic, tourists' perspectives and affections shift toward greater positivity, leading to a significant increase in tourists’ VI this destination. This finding is consistent.

This study highlights that after experiencing AI chatbots, the destination brand image is perceived more positively, resulting in a considerable rise in both tourists' VI and their readiness to pay a higher price. The research by Kani et al. (2017) and Chaulagain et al. (2019) supports this finding, emphasizing that a positive brand image can significantly enhance tourists' attitudes and behaviors toward a destination. However, our results contrast with the studies of Kock et al. (2016), Pratt and Sparks (2014), which found that SDBI had no significant effect on the likelihood of paying a premium or desire to visit. The discrepancy between our findings and theirs could be attributed to several factors, including differences in research contexts, changes in tourists' perceptions over time, along with the impact of external factors like service quality, pricing strategies and the degree of competition in the travel industry.

Moreover, to the best of our knowledge, there is still a limited number of studies that explore both key aspects of destination branding, brand image and authenticity and examine how these factors interact to influence tourists' VI and their willingness to pay more, particularly in the context of sustainable tourism. While there has been growing interest in sustainable tourism and its role in shaping consumer behavior, the intersection of brand image, authenticity and tourists' willingness to engage with and pay for such experiences remains an underexplored area.

Regarding the moderating role of nature connectedness, the results reveal unexpected findings. Although theory and previous research by Mayer and Frantz (2004) indicate that individuals with a strong connection to nature tend to show greater appreciation and support for environmental initiatives (Whitburn et al., 2019), this study's results did not find a significant moderating effect. The results show no significant moderating effect, which contrasts with the findings of Pritchard et al. (2020) and Van der Linden (2015), who found that individuals with nature connectedness are generally willing to support and invest in environmental conservation activities. NC does not moderate the relationship between brand image and behavioral intentions. This finding is contrary to the predictions based on studies by Spence et al. (2018) and Tussyadiah et al. (2018) regarding the influence of nature connectedness on emotional responses and visitation intentions. Based on the definition by Lu et al. (2005) that PI is the willingness to experiment with new technologies and concepts, the study examined its moderating effect. The moderating effect of PI was not significant on the relationship between SDBA and behavioral intentions. This differs from predictions based on the study by Hsu et al. (2023) on the role of PI in sustainable tourism acceptance. Notably, PI only showed a significant moderating effect on the relationship between SDBI and WTPP, while no significant moderating effect was found between SDBI and VI. This finding partially aligns with research by Kim and Huang (2021) and Yoo and Lee (2017) on the role of PI in enhancing the relationship between sustainable brand image and willingness to pay more. These findings suggest that while factors like nature connectedness and personal innovativeness may have a direct impact on visitor behavior, their moderating role is not as strong as theoretically predicted. Future research could focus on understanding why these moderating effects are not significant and whether there are other variables that might better explain the variability in this relationship.

This study adopts a perception-based approach, which inevitably entails certain limitations. The findings reflect tourists' subjective evaluations of sustainability-related cues communicated through AI chatbots rather than objective assessments of destinations' sustainability performance. However, this approach does not weaken the validity of the study; instead, it aligns closely with how tourists actually form destination-related judgments in real-world digital contexts. In practice, travel decisions are rarely based on comprehensive sustainability audits but are shaped by perceived brand signals, narratives, and mediated information encountered during the pre-visit stage. Consequently, the study maintains strong external validity in explaining AI-driven tourist decision-making processes, particularly in technology-mediated tourism environments where perceptions play a central role in shaping behavioral intentions.

In the discipline of sustainable tourism, our research contributes a groundbreaking approach, particularly in the application of AI technology. Specifically, we studied how information-related factors presented by AI chatbots and tourists' interactions with AI influence SDBA and SDBI, ultimately shaping tourists' behaviors, such as their VI and their WTPP for the destination.

First of all, this study is one of the earliest to expand the framework of AI chatbot information-related factors' impact and destination brand theory in the context of sustainable tourism by examining five critical information factors. Of these, three factors, including information quality (IQL), information diagnosticity (ID) and information relevance (IR), have been proven to significantly influence tourists' assessments of SDBI and authenticity. The interaction with AI chatbot also found to have the same effect, suggesting that positive emotional engagement with chatbot can enhance VI and WTPP. This not only extends the theory of emotional connection with AI technology (Kim et al., 2020; Sharma and Paço, 2025) but also enriches consumer behavior theory and technological theory within sustainability-focused tourism.

Secondly, by emphasizing their significant influence on VI and WTPP, this study clarified the role of SDBA and SDBI in shaping tourist behavior. By focusing on these elements, our research fills a gap in sustainable destination branding theories, which had not previously examined factors related to SDBA or SDBI. It also underscores a growing trend: tourists are more attentive to genuine brand commitments from destinations, moving beyond the influence of superficial “greenwashing” efforts (Kiatkawsin and Han, 2019; López-Sánchez and Pulido-Fernández, 2016; Fatma and Khan, 2024).

Thirdly, this study explored how nature connectedness (NC) and personal innovativeness (PI) moderated the relationships among SDBA, SDBI, WTPP and VI. Despite earlier studies suggesting that these factors could drive sustainable tourism behavior (Whitburn et al., 2019; Spence et al., 2018; Pritchard et al., 2020), our findings in Vietnam revealed contrasting outcomes, with no significant moderating effects identified for NC and PI. The discovery of this new theoretical insight indicates that the impact of NC and PI could differ across contexts and study subjects. It opens up new directions for subsequent research to re-evaluate how these two factors moderate the relationship.

The following implications is that our study contributes to ELM theory by clarifying how AI information-related factors can change tourist's perceptions and behaviors through dual processing pathways. Specifically, the chatbot delivers top-notch information, relevant to the tourists, and highly diagnostic. When combined with the external factor of tourists' perceptions of the chatbot, it encourages them to engage more deeply with the destination brand, which in turn affects their VI and WTPP. Additionally, the research enhances the SOR theory by explaining the relationship between stimuli (chatbot information related factor), tourists' perceptions of SDBI and SDBA (organism), and behavioral responses like VI and WTPP (response). This is one of the first studies to combine both ELM and SOR theories to explore how AI chatbot information influences tourists' behaviors in sustainable tourism, offering new insights into the influence of technological factors on consumer behavior.

Finally, by addressing a notable gap in the understanding of AI chatbots within sustainable tourism, this research diverges from the focus of previous studies, which predominantly examined information quality from sources like VR technology (Jiang and Phoong, 2024), social media (Kim et al., 2017) or studies on the intention to continue using chatbots (Pham et al., 2024), with few examining the direct role of chatbots in enhancing tourist's perceptions of destinations and stimulating subsequent behaviors. This is particularly true for situations where tourists have not yet visited a destination but are still disposed to pay additional costs for a destination that is supported by chatbot-provided information over another. By clarifying these relationships, this study contributes to the advancement of sustainable tourism theory and paves the way for new research opportunities in the development of AI chatbot technology.

The study presents several actionable implications for tourism practitioners and destination management organizations (DMOs) engaged in sustainable destination branding in the age of AI-mediated information search.

First, sustainable destination marketing organizations should strategically enhance the online visibility, consistency, and credibility of sustainability-related information across their official digital ecosystems. General-purpose AI chatbots, such as large language model–based systems, are trained on vast volumes of publicly available online content and retrieve information based on relevance, frequency and perceived authority (Ray, 2023). As a result, the sustainability narratives that these chatbots generate are highly dependent on the quality and structure of existing online destination information. To address this, practitioners should ensure that official destination websites, verified travel portals, blogs, and social media channels contain clearly articulated, up-to-date and search-engine-optimized sustainability content. More specifically, DMOs can operationalize this strategy by developing a dedicated and highly visible sustainability section on their official destination websites. This section should systematically present concrete sustainability practices, such as eco-friendly tourism activities, community-based tourism initiatives, partnerships with local suppliers, waste reduction policies, biodiversity protection programs and recognized green certifications. From a step-by-step perspective, this involves (1) auditing existing digital content to identify sustainability information gaps, (2) producing standardized sustainability narratives using clear and nontechnical language and (3) structuring this content with consistent keywords and metadata. By doing so, destinations indirectly influence how AI chatbots retrieve and synthesize information about them, thereby improving the perceived relevance and quality of sustainability-related chatbot recommendations. This process helps strengthen tourists' perceived destination brand image and authenticity, as supported by the study's findings on the role of information quality and relevance.

Second, tourism practitioners should actively facilitate guided and meaningful interactions between travelers and AI chatbots to steer user engagement toward accurate, brand-aligned sustainability narratives. Rather than leaving chatbot interactions entirely unguided, destinations can subtly shape user–AI conversations by encouraging travelers to ask specific, sustainability-focused questions. For example, official destination websites or social media pages can include prompt suggestions such as “Ask a chatbot how our destination supports coral reef restoration” or “What sustainable tourism activities are available at our destination?” This approach works by reducing users' cognitive effort during information search and simultaneously increasing the likelihood that chatbots retrieve sustainability-related content that aligns with the destination's strategic positioning. Practically, this can be implemented by embedding suggested prompts or clickable chatbot queries within destination webpages, travel planning guides or digital brochures. Such guided interactions enhance perceived enjoyment by making the information search process more engaging and interactive, while also increasing anthropomorphism by fostering conversational, human-like exchanges. Over time, these emotionally engaging interactions can strengthen tourists' attachment to the destination brand and reinforce perceptions of authenticity in sustainability communication.

Third, a more proactive and high-control strategy involves integrating AI chatbot technology directly into destinations' own digital platforms. Instead of relying solely on general-purpose chatbots, destination organizations can deploy customized conversational agents powered by large language model APIs. These systems allow organizations to train or constrain chatbot responses using proprietary and verified destination data, including sustainability policies, environmental guidelines, cultural norms and frequently asked questions (Rawashdeh et al., 2024; Google LearnLM Team, 2024). From an implementation perspective, this strategy typically involves several steps: (1) curating a verified internal knowledge base containing sustainability-related documents and policies, (2) integrating this database with an AI chatbot interface via an API and (3) configuring response boundaries so that the chatbot generates answers exclusively from approved content. This approach significantly enhances information diagnosticity by ensuring that responses are destination-specific, accurate and contextually rich. At the same time, perceived accountability is strengthened because tourists associate chatbot responses directly with the official destination organization rather than with an anonymous AI system. Furthermore, embedding such an internal chatbot within the destination's official website or mobile application enables consistent and controlled sustainability communication throughout the traveler decision journey. This internal AI assistant can also be used for staff training and internal knowledge dissemination, ensuring that sustainability messaging remains coherent across both external marketing and internal operational practices. Collectively, these measures allow destinations to leverage AI not merely as a passive information channel but as a strategic tool for reinforcing SDBI and authenticity.

Although our research provides several theoretical and managerial implications for applying AI chatbot in sustainable tourism, new approaches for later research are opened up from our article's limitations. Firstly, our research is cross-sectional as we conducted the survey in Vietnam over a period of approximately two weeks. This type of research may not allow for causal inferences between the research variables (Rindfleisch et al., 2008). As user perceptions of chatbots and sustainability-related attitudes may evolve over time, future studies could adopt longitudinal or experimental designs to track behavioral changes and improve causal inferences across different time periods and cultural contexts.

Second, the study did not collect detailed information about participants' travel background, such as travel frequency, international versus domestic experience or familiarity with different types of destinations. Although the survey introduction clearly explained the concept of sustainable destinations and included screening questions to ensure participants had experience using chatbots for travel advisory, respondents with limited travel exposure may rely on pre-existing or familiar destination images. This could influence their evaluation of chatbot support and introduce potential bias. Future research could include travel history as a control or moderating variable to examine how previous travel experience shapes chatbot-related perceptions in sustainable tourism.

Moreover, no specific AI chatbot was referred to examine its impact on tourist behaviors. In fact, there are numerous types of chatbots or virtual assistants that are deployed for various purposes such as education, banking and finance, tourism and more (Asha et al., 2023). As a result, our article may not fully capture the true insights of users when interacting with chatbots for sustainable destinations information and recommendations. Further investigations could focus on the tourist interaction with tourism-specific chatbots for the purpose of fostering more detailed insights into this area.

Third, while the survey introduction clearly stated the research focus on user interaction with content recommendation algorithms for sustainable destinations and included screening questions to ensure that only individuals who had used chatbots for travel advisory proceeded, the study did not collect detailed travel background (e.g. frequency of international vs domestic travel or familiarity with diverse destinations). Respondents with limited travel experience may rely on familiar destination images, which could influence how they evaluate chatbot support in the context of sustainable tourism. Future research could include travel history as a control or moderating variable to better understand how prior travel experience shapes perceptions.

Moreover, the selection criteria did not further separate respondents based on their familiarity or frequency of interaction, despite the survey introduction stating that the study was intended for people who had experience with AI chatbots in recommending sustainable destinations. This might have made it more challenging to accurately capture the range of users' experiences with these kinds of chatbots. In order to properly account for these variations and ensure a more realistic portrayal of real users' experiences, further study could improve the selection procedure.

Fourth, our research concentrates mainly on the contribution of the AI chatbot factors to the aspects of destination branding, without examining the sustainable destination itself, such as destination image or authenticity. Hence, future research could broaden the scope to explore the relationship between these factors. Additionally, it could investigate how other different AI chatbot characteristics affect various destination marketing aspects such as destination image, authenticity, loyalty and more.

Last but not least, due to using quantitative research methods, survey questionnaires, particularly with snowball sampling methods, there are potential limitations in the data collection, such as random answering, social desirability bias and difficulty in exploring the underlying reasons or context. Moreover, this method is limited in exploring the underlying motivations or emotional context behind user perceptions. Therefore, future researchers could adopt various methods, including anonymous interviews or experiments, which can help participants respond more honestly and accurately, thereby reducing the mentioned biases. Moreover, data from qualitative studies can provide more nuanced insights about the motivations and context behind behaviors.

The relationship between chatbot-related factors, customers' perception of sustainable destination brand characteristics and their behavioral intention has been investigated through our empirical study. Our research findings implicate many insights for both theoretical and managerial aspects. Notwithstanding the beneficial contributions, our research still has several shortcomings, which serve as the foundation for additional research efforts. For example, future studies should broaden the focus of examining the connection between chatbots and an eco-friendly travel location and look more closely at how chatbot features affect different brand marketing elements.

The research process involving human participants followed the ethical standards of the institutional and national research committees of the 1964 Helsinki Declaration and its subsequent amendments, as well as related ethical standards.

The supplementary material for this article can be found online.