The developments in artificial intelligence (AI) technology have raised concerns regarding prospects of existential risk, having inter-generational significance. While some literature on AI analyses its technology within current scope of scientific knowledge, certain others adopt a philosophical approach, synthesising abstract concepts in order to explore the very possibility of such risk. Corresponding to the latter, this paper assesses the specific possibility of existential risk posed from ethical misalignment of sub-symbolic AI, in a technologically neutral fashion. Such approach borrows ideas from psychology (particularly, cognitive sciences), philosophy, epistemology etc. so as to illuminate the essential meaning of intelligence. Presenting a fresh perspective, the paper's results show autonomy of sub-symbolic AI in a new light. Acknowledging the ‘precautionary principle’, the results are presented with a view to complement, and not contest, any established views on alignment research.