Since the famous machine v. person dichotomy has been overcame, the close relationship between technology and translation industry is getting clearer over time: artificial intelligence is positioning itself as an undeniable ally of linguistics worldwide.
How does this relationship work and how will it change in the future?
A Comprehensive Approach
Defined as a broad science field that includes not only informatics but also psychology, philosophy, linguistics, among other areas, artificial intelligence (AI) is concerned with getting computers to perform tasks that would normally require human intelligence.
In this sense, AI has different applications, such as image recognition, speech recognition, translation, Q&A or chatbots, and games. Nevertheless, its projection is huge, since it has become an integral part of our lives. From healthcare and agriculture, through energy, transportation, space exploration, education and finance, up to entertainment, environmentalism and human resources.
The greatest use of AI in product making was in the financial services industry, with over 30% of respondents using AI in 2023.
As in other industries, AI’s contribution to linguistic services is certainly disruptive. “Language is key to building winning global customer experiences. As companies expand globally, the need for accurate, efficient, scalable localization has never been more crucial. With the rise of Contextual AI and large language models, such as ChatGPT, there are exciting new opportunities to speed up the localization process and unlock significant benefits for companies of all sizes”, notes the specialized media, Slator.
Being part of such tendency implies embracing a hybrid approach, in which experimented linguists’ expertise is combined with the technological capacities of large linguistic models. “Companies can render more content in the native language of many markets while, at the same time, ensuring the accuracy of translation only a skilled linguist could provide”, Slator adds.
The future of AI-assisted translation seems to be defined by two key concepts: automation and the increase of human translator capabilities.
“Full automation means delivering raw machine translation output directly to the end user. No post-editing or any other human involvement once the machine translation system has been trained. This dates back to the early days of machine translation when the goal was FAHQT (fully automatic high-quality translation) of every source text. No attention to use case. Anything less indicates failure. However, there is another way to approach automation, an approach that adds argumentation to the mix”, states FIT North America, a regional center of the Fédération Internationale des Traducteurs.
“Global machine translation market revenue is estimated to reach USD 4,069.5 million by 2030 with a compound annual growth rate (CAGR) of 19.9% from 2022 to 2030”.
And they add: “For some jobs, the question is binary: replace humans or augment their capabilities. When it comes to professional human translators, the question is far from binary. There are many use cases. In some, such as translating post-discharge instructions for hospital patients with limited English proficiency or translating European Union legislation, augmenting the capabilities of a human translator is clearly the way to go. In other cases, such as translating items in a rapidly changing database of tech support articles or pre-trial triage of foreign-language documents for relevance, a fully automated approach has been successful, selectively followed by human translation.”
According to this new paradigm, just as “augmented reality” uses AI to enhance individual access to context relevant information, the idea of “augmented translators” provides linguists more context and guidance for their projects. Augmented translation puts the human expert at the center of the equation.
“The shift to augmented translation will not be painless, but because the technology shifts to a linguist-centered perspective, it will hold tremendous potential for those linguists who are willing to embrace the changing landscape of technology”, CSA Research predicted some years ago.
Interacting and potentiating capacities seem to be the key of the matter. The goal is to achieve technology that increases the professional human translators’ skills when full automation is too risky. To all industry stakeholders, those are the challenges to address in the future.