This paper intends to be a multi-level analytic and evaluative study of online automatic machine translation from English into Arabic and vice versa. A multi-level analysis focuses on lexical and grammatical ambiguity, syntactic structure and on semantic transfer from the source language into the target language. Meanwhile, the multi-level evaluation concentrates on the dimension of adequacy, accuracy and acceptability of this type of automatic translation. To carry out this study, six English and Arabic texts have been randomly selected and translated from the source language into the target language by using more than one of the available online machine translation systems: Al-Kafi, Al-Wafi, An-Nakel, Al-Misbar, Tarjim of Ajeeb and Google Research.