I answered on Quora:
Before you ask “why”, always ask (maybe somewhere else) “whether”. Your declaration that this doesn’t require a lot of technology is absolutely false, by a very large margin. (Question therefore downvoted.)
Automatic translation just isn’t there yet. Things like Google Translate, and the equivalent from other companies, are generally used just to get a rough idea what a document is about, to see if it’s worth bothering with a human translator. Pre-stocked translations could be much more accurate, but that would bloat the size of the software (including accompanying data), and be extremely expensive — the software creator would need to pay several professional translators to have any hope of covering even a significant fraction of the world’s most popular languages.
Either way would also drastically increase the effort needed to create the software, when it’s very often not necessary, or at least much lower priority than relevant features. That would drive up the prices of commercial software, put people off of writing open source, delay features and bugfixes, and introduce much more opportunity for bugs, including security vulnerabilities. A part of this you might not realize is the need to accommodate strings that may be much longer or shorter than in English, or written in a different direction — not just right-to-left such as Semitic languages but some also top-to-bottom, and THEN either left-to-right or right-to-left, such as some Asian languages.