Bias, "an inclination, prejudice, preference or tendency towards or against a person, group, thing, idea or belief" (Murphy 2021), raises ethical questions whether in human or machine communication and it can have detrimental impacts on individuals and societies, e.g. criminal judgments. The deployment of AI systems in real-world settings has exposed various kinds of bias against certain social groups. The big number of research on bias in AI applications generally and the few studies on bias in neural machine translation (NMT) particularly have given the present study the momentum to further investigate the issue. It aims to examine bias in NMT through exploring the translation of some heavily-loaded ideological messages from English into Arabic. A multidisciplinary perspective deriving its tenets from translation studies, political sciences and computer science is utilized to explore bias in the translation of ideology by Google Translate API. It employs a qualitative approach using analysis, comparison and deduction as tools of research. It has reached some interesting findings, which came contrary to my initial expectations.