Artificial intelligence (AI) is the science of making machines that can think like humans and using computers to do things that traditionally require human intelligence. AI embodies and invites change, hope, and opposition. It also opens up new, nearly unprecedented opportunities and troubleshoots more rapidly and imaginatively than at any time before. As the technical advancement and application speed up significantly, so does public understanding of its effects. Additionally, while mechanisms are needed to predict, control, and direct the growth of artificial intelligence in order to prevent undesirable results, it is unclear how society should define such growth considering present power relationships. In this paper, an effort is made to investigate whether it is acceptable to restrict, alter, or amend AI breakthroughs based on their potential societal, ethical, and legal implications. From the perspective that science, ethics, and politics should work towards a separation of responsibilities and a power structure instead of a conflation, the researcher critically examine four plausible justifications for stricter political and ethical supervision of research discipline. It is argued that science, ethics, and politics should be kept separate in order to preserve our ability to appropriately assess the most acceptable course of action in light of the implications of AI. We take this action out of concern that such conflation could lead to uncertain and shaky outcomes, such as politicised scientific knowledge or ethics washing, morality constricted by business or scientific interests, inadequate regulatory oversight, and political work as a result of a misguided presumption regarding manufacturing self-regulation. As a result, we suggest that in order to ensure AI research serves the society at large, the varied responsibilities of science, ethics, and governance should be honoured.