This research aims to assess the capacity of artificial intelligence–based machine translation, particularly the ChatGPT model, to convey the binding force and linguistic specificities of supply contracts between Arabic and English. The study adopts a comparative approach combining linguistic, syntactic, and semantic analysis, applied to two supply contracts arising from distinct legal contexts. The theoretical framework examines the notions of contractual obligation, condition, and exception, as well as the constraints inherent in legal translation. The empirical analysis demonstrates that, despite apparent linguistic accuracy, machine translation tends to attenuate the legal force of imperative and conditional formulations. The findings thus highlight the current limitations of machine translation in the legal field and confirm the necessity of human intervention to ensure conceptual precision and the legal effect of the translated text.