Agencies should work with competent human translators for all translations, including translations supported by translation technology.
Why create multilingual content?
All people in this country, regardless of the language they speak, deserve meaningful access to programs and activities that are conducted or supported by federal agencies.
Two executive orders affirm the federal government’s commitment to improving language access services and ensuring full participation by individuals with limited English proficiency (LEP).
- Executive Order 14091, Further Advancing Racial Equity and Support for Underserved Communities Through the Federal Government (February 2023)
- Executive Order 13166, Improving Access to Services for Persons with Limited English Proficiency (August 2000)
What is translation technology?
Translation technology is a rapidly evolving field that involves the use of computer software and other tools to support the translation process.
Computer-aided technology and
machine translation are two common approaches to language translation.
With computer-aided technology, the translation is created by a human translator with some aspects of the process aided by software. With machine translation, the translation is mostly automated.
The table below highlights other key differences between the two approaches.
|Computer-aided technology||Machine translation|
|Quality||High-quality translations.||Accurate translations for simple and routine text; may struggle with more complex or idiomatic language.|
|Cost and time||Expensive and time-consuming.||Low cost and quick; may require additional costs and time for post-editing to ensure quality and accuracy.|
|Personnel||Requires human translators who use translation software to assist in the translation process; provides a library and translation suggestions for reuse.||Mostly automated; doesn’t require human translators.|
|Customization||Customizable for specific industries, such as legal, medical, or technical translations; can incorporate specialized terminology, glossaries, and translation memories.||Little customization.|
|Examples||MateCat, SmartCat, Trados Studio, and WordFast, among others.||DeepL Translate and Google Translate, among others.|
What are the downsides of using translation technology?
There are many downsides to using translation technologies.
Automated machine translation isn’t delivering 100% accuracy yet, but it is improving. Its accuracy depends largely on the language pair you need. For example, the most-used languages, such as English, Spanish, French, German, and Mandarin Chinese, have relatively high success rates of translations due to the large amount of data available for training machine learning models. However, less common languages and language pairs may have lower success rates due to the lack of available training data.
Another downside with automated machine translations is that they can struggle with idiomatic expressions and nuanced language use — how humans actually write and speak. In turn, these issues can make the translation sound unnatural and awkward.
But could artificial intelligence (AI) make machine translation more accurate? AI — where algorithms learn human behavior and language use through neural networks — is a promising field. It has the potential of making translations more accurate as it learns, but we are still far from 100% automation.
Computer-aided technology is expensive and time-consuming now, but it can improve and lower costs as adoption grows.
Should agencies use translation technology?
Agencies should use translation technology to create accurate, findable content. But they should not rely solely on automatic machine translation services or computer-aided technology. All translations should be checked by a competent human translator.
According to the Department of Justice’s Limited English Proficiency Committee of the Title VI Interagency Working Group:
Related, the Department of Health and Human Services (HHS) tells health programs covered by Section 1557 of the Affordable Care Act:
HHS encourages the health programs to understand the strengths and weaknesses of the technology and software programs that qualified translators use.
Connect with others interested in translation technology
Want to learn more about translation technology?
Join the Multilingual Community to connect with government multilingual content managers who are working to expand and improve digital content in languages other than English.
Reach out to the National Language Service Corps, a Department of Defense program that provides language and cultural support to federal agencies. Members are highly skilled language professionals who connect, share, and grow through networking, training, and testing opportunities.
Contact the Office of Language Services, a Department of State program that provides foreign language support to the White House and other federal agencies. The program increases terminology consistency by using computer-aided translation with review by qualified language professionals and glossaries.
Visit LEP.gov/digital-services-and-websites for a wealth of resources to create multilingual content for websites and digital services, among other topics.
Check out eTranslation, the European Commission’s cutting-edge online machine translation tool. Use it to get the gist of a text, or as the starting point for a human-quality translation.
Position paper on machine translation: A clear approach to a complex topic. Clarifies in easy-to-understand terms what machine translation is and how it realistically relates to professional translation (American Translators Association, 2018)
How to avoid the pitfalls of free online translation tools. Can free online translation tools ever be of service? The answer is yes. Here are some dos and don’ts for using machine translation safely. (American Translators Association, 2022)
Machine translation. Can’t a computer do all this? Learn when to use automated machine translation and when to hire a professional. (American Translators Association, 2023)
Six challenges for neural machine translation. Despite its recent successes, neural machine translation still has to overcome various challenges, most notably performance out-of-domain and under low resource conditions. (Philipp Kohen and Rebecca Knowles in the Association for Computational Linguistics’s Proceedings of the First Workshop on Neural Machine Translation, 2017)
Understanding the societal impacts of machine translation: a critical review of the literature on medical and legal use cases. Machine translation errors pose serious risks in some environments, but there is little understanding of the nature of these risks and of the wider implications of using this technology. The article concludes that machine translation technology can exacerbate social inequalities and put certain users at greater risk. (Lucas Nunes Vieira, Minako O’Hagan & Carol O’Sullivan in Information, Communication & Society, 2020)
Quantifying the Effect of Machine Translation in a High-Quality Human Translation Production Process. This paper studies the impact of machine translation on the translation workflow a directorate-general for translation. The results show that while the machine translation leads to productivity gains on average, this is not the case for all translators. (L Macken, D Prou, and A Tezcan in Informatics, 2020)
Research on the Relations Between Machine Translation and Human Translation Machine translation ignited a reform and transformation in the field of translation. It has gone through three stages, from early dictionary-matched machine translation to corpus-based statistical computer-aided translation, and then to neural machine translation with artificial intelligence as its core technology in recent years. This paper analyzes the problems with examples and puts forward a new model of machine translation plus human translation as well as its applications. In the age of artificial intelligence, computers with deep learning ability working together with humans will produce translation works with high efficiency and better quality. (Zhaorong Zong in Journal of Physics, 2018)
All references to specific brands, products, and companies are used only for illustrative purposes and do not imply endorsement by the U.S. federal government or any federal government agency.