This research attempts to elucidate the extent of civil responsibility for damages by artificial intelligence (AI), directly or indirectly. It investigates the legal capacity of AI and its characterization within jurisprudence, preparatory to establishing how its damage should be determined, and by whom. The discussion considers various possible legal perceptions of AI, as a juridical persona, a mere machine, or like an irrational animal, within both Islamic jurisprudence and positive law. The research utilized an analytical approach by integration of induction, interpretation, reasoning, and critique. The research results showed that, within the realm of Islamic jurisprudence, the majority opinion holds that AI has no legal personality and should be considered similarly to irrational animals with regards to responsibility. Under positive law, the same is most likely to hold, unless the state steps in and gives AI juridical personality, whereby rules of liability and guarantor will then respectively follow. The research advises against the formation of specific standards for redefining responsibility with regards to AI, with stress on the necessity of specific details owing to the characteristics of AI, whereby defects, errors, and linkages to damage cause difficulty.