Article Preview
TopIntroduction
According to a report by McKinsey (Lamarre & May, 2019), the internet of things (IoT) will contribute $2.7 to $6.2 trillion to the global economy by 2025. Considering the increasing volume of the data created by the IoT, how to use artificial intelligence (AI) to leverage the data and generate more economic value has become a critical global strategy. This trend of combining AI techniques with IoT’s smart sensing capabilities is termed AIoT, which is a crucial tool for context-aware business applications. Recently, AI and IoT have become more interwoven and interdependent in many fields, such as healthcare (Zhang & Zhang, 2011), biosensing (Zou, Zhang, Huang, Bian, Jie, Willander, Cao, Wang, & Wang, 2018), traffic management, and many others. These previous studies show that the applications of AIoT can enhance not only the business operational level but improve the decision-making ability at the strategic level as well.
In the era of AIoT, since IoT devices have been widely adapted to recognizing the features for certain contexts and AI technique has focused on one specific task, it can be referred to as one form of “narrow AI” applications (Welser, Pitera, & Goldberg, 2018). In these applications, AI has empowered IoT devices to become aware of their contexts and adapt to the dynamic environment (Gu, Pung, & Zhang, 2004), thereby helping the context-aware system to operate more efficiently. More specifically, due to the potential capacity of AIoT growing with the advancements in digital technologies, the role of the context-aware systems is transformed from supportive to proactive. However, previous studies in this area have predominantly focused on the functionalities of AIoT, such as adopting AIoT to rank values for reducing routine input data (Wang, Zou, Ng, & Ng, 2017), or incorporating AIoT into recommender systems (Colombo-Mendoza, Valencia-García, Rodríguez-González, Alor-Hernández, & Samper-Zapater, 2015). Although these functional-based studies present a way to deliver improved outcomes, the key features and mechanisms for AIoT-based context-aware system are still underexplored. Moreover, AIoT-based context-aware systems are different from traditional ones because the proactive characteristic of AIoT brings a certain transformation in human work by reducing the more human effort to become aware of the practical context. This change involves a rigorous process, which needs to be aligned with the practical context, and it also includes evaluating the dynamic feature of the novel artifact, adjusting the interaction between human and machine, and thus facilitating AIoT to empower context-aware service. With this emerging phenomenon in mind, we derive our research question: “How do we design mechanisms to facilitate AIoT to empower context-aware service?”