Expert Perspectives: Quality Management for Underrepresented and Global Languages [Video]
In today’s language industry, languages are divided into haves and have-nots.
Global languages such as English get the lion’s share of investment. Professionals working in these languages have no shortage of tools and resources at their disposal.
Meanwhile, many other languages struggle for attention and resources. That means it’s hard to ensure high-quality localization—even though millions of people may rely on these languages in everyday life, or even in life-and-death situations.
How can language quality management close this gap, while providing better results for speakers of all languages? Watch the video for Beyont’s full panel discussion, featuring two experts on localization and translation QA. For more details and highlights, read on.
Meet the Panelists
Livia Florensa is a quality management and language consultant with extensive experience as a QA manager, translator, and developer of ISO standards.
Agnieszka Pepera is a senior quality manager in language services at Translators Without Borders (CLEAR Global).
Satu Suomalainen, Beyont program and project manager, guided an in-depth conversation capped off by audience questions.
Highlights: Elevating Language Quality for All
Our expert panelists delved deep into the complexities of language quality management for underrepresented languages. They also provided practical examples and insights into how their organizations are tackling this challenge.
Let’s look at some key observations and ideas worth noting.
1. What Are Underrepresented Languages?
The localization industry has built up many resources to deal with some languages. Global languages, such as English, have ample technology and resources to support high-level translation and quality management.
By contrast, many other languages lack this level of attention and investment. They may be chiefly spoken rather than written. In many cases, they are spoken by vulnerable minority groups. Some aren’t recognized as “official” languages at all.
Such underrepresented languages play a vital role in many people’s lives. Yet a lack of resources makes it difficult to provide high-quality linguistic services for their speakers and communities.
2. The Challenges for Quality Management
How can we ensure a high level of localization quality for underrepresented languages? Multiple structural barriers lie in the way.
Limited technological support: Few tools are designed for these languages, and a lack of resources makes it hard to develop new ones. This makes linguistic quality assessment a slower, more challenging task.
Shortage of skilled professionals: Many people may speak an underrepresented language. However, few may have the skills to perform high-level translation and quality assessment. Formal opportunities for training and education are often absent.
Lack of standardization: Underrepresented languages tend to lack standardized rules and vocabularies. Speakers from different regions may use different terminology, or different scripts for writing. As a result, it’s often a struggle to ensure consistency across translations.
Fundamental knowledge gaps: Basic research and information about these languages are often unavailable. In some cases, we may not even know exactly how many people speak a language or where they’re located.
3. New Developments and Solutions
Language quality management is evolving, thanks to new standards and technologies. What does this mean for underrepresented languages, as well as language quality in general?
Quality standards: The International Organization for Standardization (ISO) is leading the way on new standards for localization quality.
Recently, the ISO’s localization committee has turned to improving language quality with its forthcoming standard on evaluating translation output. Its goal is to provide standardized scoring and make linguistic quality assessments as objective as possible.
Such new standards provide stronger support for quality management across all languages, including underrepresented languages. The development process is a collective effort with contributions from many countries and regions, so the standards reflect a broad spectrum of needs and realities.
The rise of AI: Large language models such as ChatGPT work well for global languages such as English, while less well-served languages are lagging behind. Nonetheless, organizations such as CLEAR Global are developing AI resources for underrepresented languages, including new translation engines and data to support them.
Many challenges remain for AI in underrepresented languages, including a lack of financial incentives to invest in it. Up-to-date training data are often unavailable: for example, machine translation models often produce archaic outputs because they rely on Bible translations.
To create better models, researchers and developers need consistent linguistic standards—but this requires extensive legwork and consensus-building among the communities that speak those languages.
Watch Now for More Insights
The road to inclusive language services runs through quality management. Interested in going in-depth? Catch the full video discussion to learn more about ways to boost quality and deliver better localization outcomes for every language.