Do you know how Google Translate works? It uses documents, mostly from the UN, and its machine algorithm slowly learns new tidbits about X language. New languages are being added there (seriously, they just added Quechua even though Bing Translator did it first haha!). However, the accuracy of the translations isn't always spot-on. A complicated phrase like 'I hope to see what mess Google Translate makes of me.' can quickly lose its original meaning. The algorithm could possibly miss important cues like grammar, context, idioms, etc. This might be because fewer documents in low-speaker-number languages are added than those with a high speaker number.

So guess what I did? I translated 'I hope to see what mess Google Translate makes of me.' through the Maltese filter 10 times. What came out? 'I hope to see what Google Translate makes of me.' It still gives the gist of the original sentence but without the informal-ness. So I put it through the Quechua filter 10 times. What came out then? 'I’m looking forward to seeing what Google Translate does for me.' Now there is no gist! This is why we need the human-reviewed accurate language-learning stuff Duolingo uses, not the machine-reviewed sloppy messes of Google Translate. With your help, we can get Maltese on Duolingo. And maybe then we can stop getting the grammar wrong (hopefully?).