NLP is considered one of the most important areas of application in the artificial intelligence field. Natural languages are used to communicate with each other in the most common way. In a connected world, it is important to automate the translation from one language to another. Though, it is challenging to know all the languages. Therefore, it is necessary to translate one language into another through the translation process. When this translation process takes place through a machine, it is called “Machine Translation” (MT). MT is an advanced method of translation through a computer system. All MT processes face the challenge of various types of ambiguities. Ambiguity is an open challenge in the process of machine translation from the source language to target language. Here, we have addressed the most prominent NLP application, i.e. Machine Translation (MT). The most challenging issue in machine translation is the presence of various kinds of ambiguities, like lexical, syntactic, semantic, pragmatic and part-of-speech ambiguities. These ambiguities can be found at different levels of the translation process. In this paper, we have explained each type of ambiguity with examples with special focus on the English-to-Sanskrit translation. We have also discussed the significant approaches to solve the ambiguity challenges. Keywords: Ambiguity, Artificial Intelligence, Natural Language Processing, Part-of-Speech ambiguity, Word Sense Disambiguation