The following article analyses the pros and cons of AI technologies used in the field of content localization, and whether or not as AI gets more advanced it could potentially pose a threat to the human workforce or help streamline the processes.
* Guest column by Hasan Köroğlu and Prof. Ulaş Başar Gezgin
What will be the future of human translators in the face of artificial intelligence translating the texts? Alarmists grant full power to AI and state that AI will completely eliminate translation jobs. Others propose the idea that AI can’t be fully human-like, it will still require the touch of a human editor to perfect the translations.
For one thing, AI is not good at implications which corresponds to pragmatics subfield of linguistics. Pragmatics is the science of language use. If we say “it is very hot here”, our friend can open the window or turn on the air-conditioner. But we didn’t tell him to do these. Implied meaning distinguishes us from artificial semantic operators. Furthermore, translations provide a plethora of implications. Thus, it is easy to see why AI will need a human editor rather than replacing human translators.
Secondly, emotions are hard to detect by AI translators. The way we express ourselves makes a difference. Thirdly, AI translators are not good at self-narratives as they lack a self’. Fourthly, AI translators do have difficulty in translating informal ‘street language’ which is not as tidy and regular as textbook language. Related to this point, AI translators currently are unable to translate regional variations of languages.
There are other views: They claim that in a short time as short as three years, most of the business transactions will be machine-translatable (Marr, n.d.). But obviously, business transactions do not reflect the linguistic diversity of the Homo Sapiens. According to InterpretCloud (2022), AI has a long way to go when it comes to interpreting context and emotions. The argument goes: “In reality, no ultra-modern robot will ever be able to replace a human being’s spirit, his flight of thought, or his sense of humor” (InterpretCloud, 2022). According to Crystal Hues Limited (2022), “AI can never replace human translators because machines are incapable of capturing nuances that come with each language.”
The pros for AI translators are speed, self-reliance, and being free of charge; while cons are misinterpretations, allocation of wrong words, and non-sense translations (Locate Translate, n.d.). Whereas the pros for human translators are one-to-one communication, easy adjustment, and accuracy, the cons for human translators are human errors although it can be improved, cost, and slowness (Locate Translate, n.d.).
One should keep in mind that AI translation is still far from accurate (Locate Translate, n.d.). Idioms and metaphors are AI’s weak points (Locate Translate, n.d.). Nearly half of Chinese idioms are misinterpreted by AI (Shao et al., 2018). It is still risky to use AI translators for fields such as medical services (Locate Translate, n.d.). Miscommunication in medical services leads to the possibility that the patients are misdiagnosed and the result can be fatal for the patients and harmful for the reputation of the hospitals (Locate Translate, n.d.). AI translates the words, not the meanings, while human translators infer the meaning behind words and translate accordingly (Locate Translate, n.d.). Moreover, Google Translate can translate only 15% of the world languages. For the other languages, we need human translators (Locate Translate, n.d.). It is also particularly difficult for AI to translate statement like slogans (Locate Translate, n.d.).
To conclude, AI will only provide assistance to human translators and will not replace them. Human translators will find themselves in a position of editing more than translating in the near future. This will require a more educated translating human work force.
