Artificial intelligence (AI) has contributed to automation in various industries with the intention of improving efficiency and reducing labor costs. Such developments in technology have pushed organizations in both the private and public sectors to incorporate AI into their operations. According to Hengstler et al. (2016), there has been an increased application of AI in the development of medical assistance equipment and autonomous vehicles compared to other sectors.
Nonetheless, the authors note that even with advanced developments in the manufacture of AI-related products, there is skepticism in society regarding the applications of the technology. Individuals and companies alike are uncertain of the safety of these products mainly because they lack adequate knowledge on the same. Hengstler et al. (2016) suggest that the best approach to take in enhancing trust in the use of AI is viewing the relationship between the company and the technology from a human social interaction perspective. The more employees are comfortable while working with AI-enabled equipment, the higher the trust level. Consequently, the UAE needs to develop strategies that enhance the simplicity of the interaction between individuals and AI.
Building Trust in AI, Machine Learning, and Robotics
Similar to other relationships, trust between humans and the various forms of technology is hard to come by. However, Siau and Wang (2018) note that there is a difference between trust in artificial intelligence and trust in other technological inventions. Consequently, the researchers illustrate four factors that play a significant role in building the relationship between people and AI. Representation is the first component that individuals look at before deciding to use an AI. It is always important that the newly introduced technology, for instance, robots, represent human behavior as much as possible (Siau & Wang, 2018).
This approach will act as a foundation for the trust of the users. Secondly, new AI users rely on previous reviews to determine whether they are confident in technology or not, particularly if their safety is at risk. Thirdly, the researchers also indicate that transparency and the ability to understand the functions of an AI are crucial in developing trust. The technology should be capable of justifying the decisions its making and the resulting behaviors. Finally, usability and reliability are essential in the continuous trust of artificial intelligence, and it is essential that the AI is designed in a manner that is easy to operate. The four suggested factors must be integrated into the UAE strategies if citizens are to trust the use of AI successfully.
Cloud-Based Life-Cycle Management for Trusted AI
Trust in the use of AI in organizational setups transcends building it in the primary stages. Companies need to ensure that the element of trust is maintained throughout all operations. Consequently, Hummer et al. (2019) suggest a cloud-based life-cycle model to ensure organizations are ready for AI adaption and leverage its full potential in its application. ModelOps is an example of a framework that companies can use to use the technology is used effectively in various operations. The algorithms used in the model consider the needs of the environment, which includes the employees in ensuring that everyone can understand the AI functionalities (Hummer et al., 2019). Therefore, it is necessary for UAE to select AI application frameworks that enhance the attitude and perception of users of the technology.
Ethics Guidelines for Trustworthy AI
Trust is categorized as an ethical standard or value in most organizations, and therefore artificial intelligence needs to be addressed with the same level of seriousness for effective implementation. According to AI HLEG (2019), it is necessary to have ethical guidelines in the introduction and implementation of artificial intelligence in an organizational setup because the guidelines play a great role in enhancing employee confidence. Consequently, a trustworthy AI comprises three distinct elements that must be integrated throughout the technology’s life cycle.
First, the AI and its functionalities should be lawful and adhere to related standards and regulations. Secondly, the management should ensure that it is morally acceptable and complies with relevant ethical values and standards (AI HLEG, 2019). Finally, the AI must be robust socially and technically to avoid any physical or emotional injuries that can be caused by the innovation. The three components comprise a framework that the UAE should apply in convincing users or employees that the AI in use safe and beneficial to their work routines.
Trust Variable: Our Framework
Trust in “Think AI” Workshop and Trust as Thesis Moderator
The choice of trust as a moderator in the research is due to the fact that it is a significant determinant in the readiness of organizations in the UAE to incorporate AI into their operations. Consequently, there is a link between the results of the “Think AI” workshop relating to trust and its application as a variable in the thesis. For instance, regulation and trust are interconnected, and this is because rules guide the use of the AI and therefore stipulates the necessary precautions to be taken while using the technology. Consequently, attaching regulations to trust as a moderator demonstrates a potential change in perception towards AI from employees.
Similarly, AI users feel safe when there are standards do guide their interactions with artificial intelligence. Moreover, in using the Technology, Organization, and Environment (TOE) model, standards in AI cut across the three aspects. The absence of standards in the three elements will imply that the UAE is not ready, which interferes with the level of trust among potential AI users and, in turn, affects the technology’s adoption.
Increasing Trust in AI Services
The technical aspect of the TOE framework used in the thesis is essential in determining whether the UAE has the right technology and expertise to handle AI. The safety of artificial intelligence originates from its manufacturers, and it is necessary that companies collaborate with credible AI developers. As discussed in the previous section, consumers highly depend on previous reviews to decide whether to embrace a technological invention or not.
Therefore organizations must ensure their manufacturers are trustworthy to increase the confidence of the specific AI’s use among its employees. According to (Arnold et al., 2019), most artificial intelligence manufacturers use supplier’s declarations of conformity (SDoC) to illustrate the product’s lineage as a way of assuring customers of the AI’s safety. In relation to the thesis, the confidence of the UAE organizations in a particular AI company results in increased trust among employees in artificial intelligence. When the organization is AI-ready, then trust can easily be enhanced since the technology in question has been marked as safe for use.
The Two-Dimensional Approach
Similar to team relationships in an organization that require transparency among the members, trust is also a requirement in the relationship between humans and artificial intelligence, and that is why it has been selected as a moderator. However, Sethumadhavan (2018) indicates that trust in AI can be evaluated in two-dimension to distinguish the role of trust and distrust in preparing an organization for the adoption of artificial intelligence. The author illustrates that while trust demonstrates feelings of confidence and safety, distrust, on the other hand, represents worry and fear in using the AI. Consequently, the two elements are crucial in justifying trust as a moderator in studying the preparedness of the UAE to fully integrate artificial intelligence in its industries (Sethumadhavan, 2018). Moreover, trust is a complex human characteristic in technology use defined by several other factors such as age, gender, culture, and personality.
All these components must be understood by organizations, in this case, the UAE government, to successfully implement artificial intelligence. Organizations need to focus on understanding the causes of distrust and mitigate them to ensure the users feel secure while using AI.
Trust: A Two-Way Traffic in AI Implementation
As discussed earlier, it takes a higher level of convincing for an individual to trust a product entirely, and this also applies to artificial intelligence. Besides trust, other challenges identified in the “Think AI” workshop included the lack of appropriate talent and poor understanding of artificial intelligence. According to Duffy (2016), it becomes difficult for employees or consumers to embrace a technology they have little knowledge about and no skill to help them use it. Consequently, most users turn to the internet, which might be confusing considering there are loads of information pertaining to the use of AI.
Trust is a perfect variable for the research because all AI preparations, including training of current and future employees, need to be linked with trust (Duffy, 2016). In this context, the UAE will only be successful with artificial intelligence acquisition in the workforce if the citizens are educated directly by the government to give them the confidence in using AI. If organizations expect users to trust AI applications, they need to understand the technologies’ functions and benefits.
References
AI HLEG. (2019). Ethics guidelines for trustworthy AI (pp. 2-24). European Union.
Arnold, M., Piorkowski, D., Reimer, D., Richards, J., Tsay, J., & Varshney, K. et al. (2019). Factsheets: Increasing trust in AI services through supplier’s declarations of conformity. IBM Journal of Research and Development, 63(4/5), 6:1-6:13. Web.
Duffy, A. (2016). Trusting me, trusting you: Evaluating three forms of trust on an information-rich consumer review website. Journal of Consumer Behavior, 16(3), 212-220. Web.
Hengstler, M., Enkel, E., & Duelli, S. (2016). Applied artificial intelligence and trust—the case of autonomous vehicles and medical assistance devices. Technological Forecasting and Social Change, 105, 105-120. Web.
Hummer, W., Muthusamy, V., Rausch, T., Dube, P., El Maghraoui, K., Murthi, A., & Oum, P. (2019). ModelOps: Cloud-based life-cycle management for reliable and trusted AI. 2019 IEEE International Conference on Cloud Engineering (IC2E), 113-119. Web.
Sethumadhavan, A. (2018). Trust in artificial intelligence. Ergonomics in Design: The Quarterly of Human Factors Applications, 27(2), 34-34. Web.
Siau, K., & Wang, W. (2018). Building trust in artificial intelligence, machine learning, and robotics. Research Gate, 26-40. Web.