05 October 2022
Guest coloumn

How to survive the storm of AI in translation?

A ‘no fear’ guide to professional translators and LSPs

Professional translators, especially in India and other countries with rapidly expanding language services markets, tend to be skeptical of and confused by machine translation. The worry is that machine translation will cut into their income now and eventually eliminate their jobs.

Are we making progress or losing ground, they wonder. Today, AI seems to instill the same feelings of loss and dread across nearly the entire business world. The fact that machines are replacing human labour is an undeniable reality. With each passing day, machine translation systems improve in intelligence. Is it possible that current jobs, such as post-editing, will be eliminated in the future?

Despite my 20+ years of experience as a professional translator and LSP founder, I admit that I am afraid. Our business has already been disrupted by machine translation. With the introduction of neural machine translation in 2017, MT in Indian languages has become a reality. Google, Microsoft, Amazon, Meta, and other companies have already left their mark on the major Indian languages. After conquering the major ones, they advanced to the next set of languages such as Tulu and Bhojpuri. Dogri and Mizo.

Have you heard the name of Meta's (Facebook's) new project? It reads fantastically, "No Language Left Behind"! Yes, they want all of the world's languages to be spoken in the metaverse they are creating. How? AI language generation, machine translation in text, voice, audio, and video, AR/VR, and so on. Every language is an oil well if data is new oil. As AI approaches its goal of conquering languages, a brave new world is taking shape in which everything we know about the language industry today will be forever altered. Is my job safe in this paradigm shift?

Our poor LSPs and translators are confused. What should be done if we are to exist in the industry at all, let alone profit?

In the early days of the AI era, predicting is dangerous in a fast-changing and unpredictable world. However, we can clearly see some industry trajectories. Technology may be changing overnight, but only in the labs and in public relations. In reality, there is always a period of transition. Machines require humans desperately during this transformative stage.

Consider the simple case of text translation: MT works its magic, and texts are translated in real time. However, in the real world, any content that is seriously considered reliable and/or publishable requires humans to be involved. Professional human translators and editors with strong translation, copy-editing, and proof-reading skills are required.

The collaboration between man and machine is then divided into three stages: pre-translation, translation, and post-translation. Professionals are required for pre-editing the source content, creating styles, and building glossaries for large and branded content. As previously stated, we require human editors to post-edit the MT output for accuracy, fluency, quality, and, more importantly, brand consistency. All quality checking and integrational tasks must be completed after translation and editing.

Even for short-lived content such as news or social media posts, branding and accuracy are non-negotiable. Style and glossaries are unavoidable in larger projects. Even if the machines are fine-tuned and re-trained with human-perfected copies, their contribution will never be completely eliminated. Focus will shift from ‘number of words’ to ‘overall contribution’.

As a result, good post-editors are in high demand. In fact, even terms like post-editing and post-editors are in flux. Today, we simply call them "language experts." These specialists will collaborate with the MT or multilingual text generation environment at all stages of content processing and management.

Furthermore, as MT lowers the cost of translation and AI makes it available to anyone at a lower cost, the customer base is expanding in every direction. The future unquestionably belongs to multilingual content makers, and monolingual content makers would have to be satisfied with the local market or perish. The race has already begun, and every company worth its salt is developing their own multi - language strategies. There will be no language left behind because the companies do not want any market to go untapped.

No language professional will be left behind too, because the language game is and will continue to be a man-machine combination. You only need to know how to win the game as an LSP or freelancer. The major decision to stay alive is to improve one's skills. More than that, understanding and accepting change is critical because it is how evolution works.

—---------

The author is the founder of Ailaysa, a platform for AI translation and multilingual content. He is the CEO of Langscape Language Solutions Pvt Ltd and the former Vice President of Citlob (South). Send an email to senthil.nathan@ailaysa.com.

Senthil Nathan

Founder | Ailaysa