Keywords

1 Background

While studies are proliferating on how AI may change the face of tourism [6, 10, 11], there’s an obvious lack of describing which specific AI technologies are in question and how they particularly relate to tourism. This situation risks the use of AI as yet another hollow ‘buzzword’. AI is indeed a game changer for tourism [7] – but we need to specify which technologies we mean by AI, how they work and what they mean for tourism. There are many sub-branches of AI, such as machine learning and deep learning, and delineating them is necessary for maximum utility [12, 13].

While all technologies that aim to mimic human intelligence in non-human platforms can be labeled as Artificial Intelligence (AI) [14, 15], machine learning (ML) refers to a subset of these technologies which comprises software applications that are able to learn to predict outcomes of actions or inputs without human intervention [11]. Machine learning can broadly be categorized into three subsections: supervised, unsupervised, and reinforced learning [16]. Among these categories, multi-layered algorithms that are modeled after the neurons in human brain (artificial neural networks) to make more complex decisions collectively make up the deep learning subfield [17, 18]. Deep learning (DL) is believed to embody the farthest advancement in AI technologies as it creates models that are able to learn from complex environments and make optimal decisions without the need of human input [19, 20]. Therefore, this study focuses on emerging AI-based technologies rooted in deep learning where potential breakthroughs in service provision and customer satisfaction exist.

2 Purpose of the Study

The purpose of this propositional study is to a) identify and review emerging AI-based technologies, chiefly rooted in deep learning, that have a high potential of operational utility for tourism industries, b) propose specific application areas for each identified deep learning technologies within the context of tourism and hospitality.

3 Review of AI-Based Technologies and Application Areas

Upon review of AI, ML and DL literature as well as industry reports, seven AI-based deep learning technologies identified and reviewed in this study were Convolutional Neural Network [20], Style Transfer, Deep Learning Based Recommendation System [21, 22], Generative Adversarial Network, Variational Autoencoder, Recurrent Neural Network and Graph Neural Networks (GNN) [23, 24]. Definition, function, and potential application areas of each technology are summarized in the Table 1.

Table 1. AI-based DL technologies and potential application areas in tourism

4 Conclusion and Future Implications

This review appraises the most prominent deep learning technologies with applicability to tourism and hospitality industries. Two important highlights of this review were 1) the indispensability of interdisciplinary frameworks to study the utility of AI-based technologies in tourism, and 2) the challenge of reliable data in tourism and hospitality domains. A crucial aspect of AI-based technologies is that they are highly data-dependent. This is a challenge for potential AI applications as sustainable solutions to adequate and accurate data collection are lacking in many tourism sectors. Therefore, it might be necessary to prioritize operational areas that are more conducive to reliable data collection than others such as international border crossings, hotel customer registrations, ticket sales for attractions, etc. AI-based applications may be more likely to succeed if they are first applied in these areas and then expand into other contact areas as reliable data linkages are established.