This study investigated how Internet users construct a sense of ambiguity in the English language to make funny jokes through Internet memes. Its aim was to examine the ambiguity types which stimulated humor in internet jokes. The jokes investigated were images presented with puns that were shared in public through an online platform called 9GAG. The research method applied was the qualitative method. The data, consisting of 50 jokes, were selected from the online 9GAG platform and listed by their categories. The researchers analyzed the categorized jokes using an observation table and drew up the formula of each sentence based on the categories identified. The categorization of ambiguities in this research was decided based on the semantic theory of ambiguity, which consisted of lexical, referential and syntactic ambiguities. The results showed that lexical ambiguity dominated the popularity of language construction used as ambiguous jokes, numbering 29 jokes (56%). It is followed by 14 (30%) syntactic ambiguities and seven (14%) referential ambiguities. The results revealed that lexical ambiguities involving the literal and figurative meanings of homonyms were frequently used to make jokes in the Internet memes. Such ambiguities could produce a humorous effect and cause some confusion about the language used. However, the study had limitations, one of which was the limited use of samples. Future research into lexical ambiguities should considerably increase the number of samples.