At the 21st International Federation of Translators World Congress in Brisbane from 3-5 August 2017 there was a lot of talk about neural machine translation and artificial intelligence (AI) and what kind of threat this presented to translators and interpreters. Overall, the message that came through was to embrace technology to improve productivity and accuracy and to rest assured that more data won’t necessarily deal with the complexity of communication and layers of language such as syntax, semantics and pragmatics.
What is all this talk about AI?
Good question. Artificial intelligence (AI) is the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. A couple of words of AI jargon that you’ll probably start hearing more frequently…
AI complete: the most difficult problems are informally known as AI-complete or AI-hard, implying that the difficulty of these computational problems is equivalent to that of solving the central artificial intelligence problem—making computers as intelligent as people, or strong.
Singularity: the hypothesis that the invention of artificial superintelligence will abruptly trigger runaway technological growth.
So translation is still currently considered AI complete.
If you are interested in the impact of AI on all professions in general (especially legal and medical) I recommend the AI Race produced by the ABC.
What about advances in technology?
It is true that despite the technological advances and more contextually sensitive machine translation machine translation still produces unacceptably poor quality content, especially companies and brands that set a very high bar for their content and brand voice. The continuing improvements to machine translation will however have an impact of the role of humans in the translation process.
We need to emphasise the work of the human brain or as it can also be referred to, wetware (human brain cells or thought processes regarded as analogous to, or in contrast with, computer systems). Wetware is the source of contextual understanding, flow and style that a machine cannot produce to the same extent. It is capable of the creativity, understanding, and personality that make for truly effective translation, localisation or transcreation.
Also, don’t forget that there are still languages that google translate doesn’t understand and if you want to be assured that there is significant disparity between hype and reality, just take a look at the nonsensical results of Google Translate Sings (e.g. “Hello” by Adele),