The concept of the "native speaker" (NS) remains a constant interest in the field of language pedagogy due to the need the field has for models, goals and norms (Davies, 2003, p. 1). On the other hand, some other researchers such as Byram (1997) in his model of intercultural competence (ICC) and Cook (2012) in his model of multi-competence have argued against the NS as a norm of assessment and deemed it an idealistic, unrealistic goal. In the context of Arabic language, given its complex sociolinguistic situation, it is essential to identify the role of NS for assessing the non-native learner's level of proficiency. This paper discusses the usefulness of the NS as the only source of language data, and investigates whether it should be taken as an assessment norm in an Arabic language context. Semi-structured interviews were conducted with nine assessment stakeholders who fall into three categories; assessor teachers, teachers, and learners of Arabic as a foreign language in the UK Higher Education context. Findings showed contested opinions regarding the assessment norm of Communicative Arabic Proficiency (CAP), and pointed out an existing norm of the educated native speaker despite the complex relationship between the concepts "proficiency and education" in the Arabic context. Alternatively, a diglossic competent L2 user was identified as a potential norm, although the definition of this concept remains unclear.