Is Google going to reject machine translations?
In recent remarks, lawyer and Google representative John Mueller said that content generated automatically with artificial intelligence tools went against the company’s guidelines.
You can see the guidelines on AI content in Google’s official documentation, in which it reserves the right to remove content with certain characteristics that show signs of being based on artificial intelligence: apparently random text containing SEO key words or contents clearly taken automatically from existing blogs or websites (a process known as “scraping”).
All this has implications for machine translation. A lot of translation work today uses software and algorithms that generate computer-assisted translation, which is then revised and edited by professionals in post-editing processes. So this raises an important question. Throughout the translation service industry, if AI-generated contents are labelled as spam, will the company eventually include machine translation in this category? This doesn’t seem to be the case if you look at Google’s official guidelines. There are two reasons for reaching this conclusion:
- It would be wrong to automatically presume that the original text (before machine translation) was generated automatically
- In a professional context, all machine translated content must undergo post-editing
Be that as it may, the capacities of the search engine’s algorithms are still a little obscure and unpredictable. Can Google really detect this type of content efficiently and without discriminating against translated content? It is a legitimate concern on the part of professionals in the language industry that they might see their translation, localisation, post-editing and transcreation work set aside. It is therefore essential to see how Google’s position on the issue will evolve over time.