Translation ethics and machine translation
In our last article, we explored how the global demand for translation drives language technology development. One of the most common technologies, machine translation, is growing more and more powerful. It’s time that we stop to look at translation ethics and ask: is machine translation ethical?
Before we answer that question, let’s look at Google Translate, a statistical machine translation (SMT) tool. Users of the tool must agree to Google’s terms of service:
“When you upload, submit, store, send or receive content to or through our Services, you give Google (and those we work with) a worldwide licence to use, host, store, reproduce, modify, create derivative works (such as those resulting from translations, adaptations or other changes that we make so that your content works better with our Services), communicate, publish, publicly perform, publicly display and distribute such content. The rights that you grant in this licence are for the limited purpose of operating, promoting and improving our Services, and to develop new ones. This licence continues even if you stop using our Services (for example, for a business listing that you have added to Google Maps).
In other words, Google can use any data that you enter into Google Translate. Even when you stop using the service, the license continues. Your text now belongs to Google and they can use it worldwide and free of charge.
Breaching non-disclosure agreements
Many companies have non-disclosure agreements to protect their intellectual property, brand and confidentiality. This means that a translator cannot share the text with a third party. But when the translator uses a machine translator like Google Translate, the translator unwittingly shares confidential information with a third party. Obviously, this is a problem not only for translation ethics, but legality as a whole.
Proper translator compensation
In addition to sharing confidential client information, machine translation robs translators of their own expertise. Drugan and Babych (2010) bring up the ethical concern that translators lose the ability to be properly recognized and compensated for their work.
Translators who are the first to translate a technical term, for example, often receive recognition. But if the translator uses machine translation, their work is freely available to others. The translator has no control over consent, and receives neither credit nor proper compensation.
Quality of the product
In addition to confidentiality concerns, SMT tools like Google Translate simply generate lower quality work. Since SMT tools run on large collections of data, the results for uncommon or complicated text are often messy or downright wrong.
So, translators who use SMT may be able to output more words per hour, but the quality goes down. However, most translators charge a fee for a human translation. Both client and translator lose here. The client loses a fee paid in good faith for a human translation, and the translator loses their reputation as quality goes down.
Red flags for translation ethics
When you look at the entire picture, the huge ethical concerns that machine translation brings to the industry cannot be ignored. A translator who breaches non-disclosure agreements, shares intellectual property and consistently delivers low quality work harms not only the individual translator, but the industry as a whole.
Yes, the industry is rapidly growing. Automation is tempting. Yet, when considering machine translation ethics, it’s important for the industry to be alert and mindful of the consequences.
This article is based on an academic study by Kaori Myatt, titled “The ethics of machine translation and crowdsourcing.”