In a world in which IT is developing faster than ever, providing reliable solutions to all problems, regardless the field of interest, the issue of computer-assisted translation systems is more and more complex, offering both advantages and disadvantages. The field of translating legal texts from Romanian into English is deprived of a large number of specialists and this is one of the main reasons why more and more people resort to computer-assisted translations, especially Google translations when dealing with translating various texts (from Romanian into English or vice-versa), in this particular case legal texts.
Yet, although this field has immensely thrived recently, and is quite reliable for simple and very easy translations, the average English speaker runs huge risks of falling into traps that can lead to errors and misinterpretations. Thus, this article aims to identify the main theoretical approaches to computer-assisted theories and the major risks and threats that occur in this type of translation, focusing particularly on legal texts. There are important differences, not only of words, but also differences in the legal systems themselves, both theoretical and practical ones, which must be correctly dealt with –something which computer-assisted translation cannot do (yet).