The Impact of AI on Translation | Argo Translation

There is no hotter topic in our industry than AI's potential impact on translation. The pessimists in the crowd say, "That's it, pack it up. Translators will lose their jobs, and interpreters are no longer necessary. This is the end of our industry." These fears are not new. Similarly, the advent of EDP, or Electronic Data Processing, rose to prominence in the 1950s. Most large corporations in America focused on how to automate the regular tasks involved with processing large amounts of data.

Of course, technology presented a massive shift in the workday for your average office worker. The change in work offered vastly different opinions. Does this sound familiar? What's old is new again. When we started our company almost 30 years ago, many people suggested that Machine Translation would make us redundant. Since then, the industry has only grown by leaps and bounds. Projections show the industry to grow by 5% annually through 2025, reaching 77 billion USD in global sales.

The most significant impact of Artificial Intelligence (AI) and Machine Translation (MT) on our industry is that translation and interpretation services are now available to a broader audience. The costs and risks associated with starting a translation initiative have dropped. The introduction of AI has created more opportunities for translation companies, not less! The technology has allowed us to tailor solutions that involve various levels of services bringing human and AI resources to the table. The most significant risk for a language services buyer is a potential lack of transparency. Many unscrupulous language providers are weaving AI solutions into their workflows and not disclosing them to their customers. The massive growth in our industry backs up this assertion.

So, how do I navigate this new world of language services if I am a buyer? How do I understand the differences in service? How do I manage the relationship with my language service provider to maximize my investments?

Different levels of service in language services

It is essential to understand that all translation isn't created equal. As a buyer of translation services, you must maximize your investment. You should never pay human translation rates for machine translation services. This mistake is incredibly common today. Many language service providers (LSP) simply weave machine translation into their workflows and don't disclose it to their clients.

Simply put, this is dishonest and dangerous. The LSP should never judge what level of risk a client is willing to accept with their translations. That principle is the bottom line here. LSPs in good faith must be transparent about the pedigree of their work with end clients, period.

As a translation buyer, you should know about the different levels of service typically offered by LSPs:

• Machine translation (MT) – straight machine translation is where content is translated using MT technology to create the translation without human review or oversight.
• Machine translation plus post-editing (MT + PE) – content created by machine translation is reviewed for errors by a human editor (i.e. MT-boosted).
• Human translation – content is translated by a human translator who reviews their work.
• Human translation plus independent editing – content is translated by a human translator and then reviewed by an independent editor.

The bottom line is that you should know precisely what you are paying for, to compare a proposal to others properly. You also should know about the workflow to assign the appropriate amount of risk to the project. If you use an AI or MT-boosted workflow, you might transfer additional internal QA or even place a waiver at the bottom of the document. AI can work well for specific use cases, but transparency is required for adequately handling the associated risks.

Managing the relationship with your language service provider to maximize your investment

Communication is critical to a productive relationship with your translation service provider. When kicking off a new initiative or project, it is imperative to cover the following topics.

• Importance of project – be sure to explain the importance of the quality level of your project. Your translation provider will always try to get the content 100% correct, but different service levels typically feature different levels of quality assurance. You would never use machine translation for an important message to employees from the CEO. You would never just use machine translation for a user's manual for a defibrillator. A clear understanding of the acceptable level of oversight for each project is a critical component of proper translation.
• Budget – candidly speak with your translation supplier about your budget. Let them know if you have a limited budget so they can explore different service levels. A project's oversight and quality assurance level _­will directly impact the budget.
• Timing – if you have a large volume project that requires completion in a short time frame, certain levels of automation can positively impact the time necessary to finish the translation.

All three of these factors must be balanced to develop the appropriate workflow and process for your project.

Final thoughts

Responsible use of AI and MT will positively impact the future of the language services industry with improvements in timing and pricing for translation projects. Proper use of technology will also pave the way to more engagement across all types of use cases. The reduced cost and time required for translation services will only usher in an expanded demand for language-related services. This outlook will also develop the need for the role of human translators in multiple steps in the process. Quoting my favorite ride at Walt Disney World, The Carousel of Progress, "There is a great big, beautiful tomorrow, shining at the end of every day."