Investigating the Role of Artificial Intelligence Voice Assistants for Arabic-Speaking Users within the Context of Knowledge-Based Systems

Artificial Intelligence Voice Assistants (AIVAs) are becoming more and more part of our daily life, but its usage in Arabic-speaking societies is not evenly distributed due to diglossia, code-switching, culturally specific communication patterns, and the lack of its integration with organisational knowledge-based systems (KBS). Weaknesses in dialect recognition, mixed-language interaction and lack of culturally based evaluation frameworks to use in real-world organisations have also limited previous studies on Arabic AIVAs. This paper examined the influence of AI voice assistants on Arabic-speaking users in the United Arab Emirates and looked at the relationships between perceived benefits, challenges, design preferences, effective usage and government communication and the perceived role of AIVAs. The survey design was a two-round cross-sectional survey design. The first round investigated usage patterns, language preferences, cultural issues, technical knowledge, and user expectations, and the second round evaluating an integrated conceptual model with structural equation modelling.

The results showed that the last model had a very high fit on the survey data and accounted a large percentage of variance on the perceptions of respondents about the importance of AIVA. The respondents indicated that they were highly exposed to AI voice assistants but less frequently used in Arabic. The primary perceived barriers related to dialect recognition, code-switching, privacy, transparency, cultural appropriateness and limited integration with organisational systems. The results also indicated that reported benefits, perceived challenges, design preferences, and communication in government significantly correlated with the perceived significance of AI services. The paper adds a culturally and linguistically based list of requirements of Arabic AIVAs, such as dialect sensitive interaction, code-switching support, privacy-conscious consent, culturally sensitive communication, accessibility, and transparency of the source. It also suggests a conceptual blueprint of AIVA-KBS integration that would enable the Arabic-speaking users to obtain authorised organisational knowledge in a controlled voice interface. This blueprint is a proposed future system, not a technically implemented or tested system. Altogether, the research is a theoretical, empirical, methodological, practical, and policy related advice on how to create more inclusive and trustworthy Arabic AIVAs in an organisational and public-service environment.