The Modern Translation Landscape: Services and Challenges

Translation services have long been one of the most demanding and rewarding offerings of libraries and information centres. As trade became global, research grew international, and the internet reached every corner, the need to move knowledge across languages multiplied. Yet the way translation services are organised today looks very different from the system that existed a few decades ago. The big national translation centres that once acted as central hubs have largely faded, private online agencies have surged forward, and artificial intelligence has reshaped what a translator’s job even means. This post walks through the present scenario of translation services, the institutions that have risen and fallen, and the tensions that define the industry now.

The decline of national translation centres 🔗

For much of the twentieth century, large centralised organisations handled the collection, processing, and supply of translations, especially scientific and technical material. These centres acted as clearing houses. They gathered translations made by various bodies, kept registers of what had already been translated to avoid duplication, and supplied copies on demand. Two names stand out in this history: the National Translation Center in the United States and the International Translations Centre in Delft, Netherlands.

The National Translation Center began in 1953 as the SLA (Special Libraries Association) Translation Pool and was housed at the John Crerar Library in Chicago. It collected and processed translations from Western European and Oriental languages, and its well-known publication, Translations Register-Index, ran from 1967. The collection eventually became part of the Library of Congress in 1989. However, the Library of Congress closed the National Translation Center, ending a long chapter of centralised translation document delivery.

Why did such established institutions wind down? The reasons are practical rather than mysterious.

Funding cuts and shifting priorities 🔗

Reduced funding for information services was the single biggest pressure. Maintaining a physical pool of translations, cataloguing them, and supplying copies required steady budgets and dedicated staff. As institutions trimmed their spending on ad hoc translation work, the volume of translations passing through these centres dropped sharply. The same financial squeeze contributed to the closure of the International Translations Centre and the discontinuation of its prestigious publication, the World Transindex.

Technology made central pools redundant 🔗

The second reason was technological. The whole point of a translation pool was to avoid duplicating expensive translation work. If a journal article had already been translated for someone, the register let the next researcher locate that translation instead of paying for it again. But the spread of the internet, online databases, and electronic document delivery weakened this logic. Researchers could increasingly find and obtain material directly. Meanwhile, the development of computer-based translation tools, such as multilingual dictionaries, terminological databanks, and machine translation systems, offered faster routes to a usable translation. The centralised pool model simply could not compete with on-demand digital access.

It is worth noting that translation centres did not disappear entirely. Specialised bodies continue to exist, often serving niche needs. The National Virtual Translation Center, established in 2003 under the FBI, provides translations in over 120 languages for the United States intelligence community. The model shifted from a public document-delivery pool toward focused, often government-linked, service providers.

The rise of online translation services 🔗

As the centralised model receded, the gap was filled by a fast-growing ecosystem of private online translation agencies. These companies operate over the web, accept documents electronically, and deliver multilingual translations on tight schedules. The convenience is obvious: a business in one country can upload a contract and receive a translated, formatted version within hours or days, in dozens of language pairs.

This shift mirrors a broader market reality. The global language services industry is now worth roughly 72 billion dollars, with the machine translation segment alone projected to approach a billion dollars within the next decade. Demand keeps rising because companies want to sell, communicate, and operate across borders, and customers prefer content in their own language.

Why demand for localisation keeps climbing 🔗

A major driver is what the industry calls localisation, which means adapting content not just linguistically but culturally for a specific market. Research cited across the industry shows that a large majority of online shoppers are more willing to buy when product information appears in their native language. For a country like ours, with more than twenty official languages and enormous linguistic diversity, this is especially significant. Businesses, e-commerce platforms, streaming services, and government portals all need content rendered into multiple regional languages to reach their full audience.

Services beyond plain translation 🔗

Modern agencies rarely offer translation alone. They bundle related services such as subtitling for films and video, transcription, interpretation for live events, terminology management, and the building of translation memory files that store previously translated segments for reuse. This packaging reflects how translation has moved from a one-off academic favour into a structured commercial industry with workflows, quality checks, and software platforms.

UNESCO’s Index Translationum and global bibliographic control 🔗

While commercial agencies handle the day-to-day flow of translation, the task of keeping a global record of what has been translated falls to a different kind of effort: bibliographic control. The most important tool here is UNESCO’s Index Translationum, an international bibliography of translated books.

The Index Translationum was created in 1932, originally under the League of Nations, and was later taken over by UNESCO. Its purpose is straightforward but ambitious: to maintain a comprehensive list of books translated and published around the world, across every discipline, including literature, social sciences, natural sciences, art, and history.

How the database works 🔗

The system depends on international cooperation. Every year, the national libraries or bibliography centres of participating countries send UNESCO data about the books translated and published in their territory. UNESCO compiles this into a single searchable database. Entries before 1979 exist in the printed editions held in national depository libraries and at the UNESCO library in Paris. From 1979 onward, the records were computerised, and the online database now holds over two million entries drawn from roughly one hundred UNESCO member states.

Why bibliographic control matters 🔗

The value of such a record goes beyond simple listing. It supports research and analysis about the global flow of ideas: which languages are most often translated from, which countries publish the most translations, and how cultural exchange has evolved over time. For librarians, scholars, and policymakers, it answers questions like whether a particular work has already been translated into a given language. In effect, the Index Translationum performs at the global scale what the old national translation registers did locally, recording and organising translation activity so it is visible and traceable. That said, the database has faced criticism over delays in updating, which reflects the difficulty of keeping such an enormous international effort current.

The Indian effort: National Translation Mission 🔗

It is useful to look at how this plays out closer to home, because the present scenario is not only about decline and privatisation. The National Translation Mission (NTM) is a government scheme designed to establish translation as an industry and to make knowledge texts accessible in all the languages listed in the Eighth Schedule of the Constitution. The idea was first proposed by then Prime Minister Manmohan Singh during the inaugural meeting of the National Knowledge Commission, which was chaired by Sam Pitroda. The mission was formally launched in 2008.

The NTM is implemented through the Central Institute of Indian Languages (CIIL) in Mysore, which acts as the nodal agency. Its core goal is to translate higher education textbooks, most of which are available only in English, into the 22 scheduled languages, so that students and academics can learn key subjects in their own language. The mission also works on standardising technical terminology in collaboration with the Commission for Scientific and Technical Terminology, and it has produced bilingual dictionaries and translated knowledge texts across many disciplines. This shows that while the old document-delivery pools faded, the underlying need for organised, publicly supported translation has not disappeared; it has simply taken new institutional forms suited to national priorities.

Challenges in the translation industry 🔗

The defining tension in translation today is the relationship between automation and human expertise. Neural machine translation, powered by tools like Google Translate, DeepL, and large language models, has advanced dramatically. It is fast, inexpensive, and capable of handling enormous volumes of text. This has genuine benefits, but it has also created serious challenges.

Automation versus human expertise 🔗

Machine translation excels at speed and cost but struggles with what humans handle naturally: cultural nuance, tone, ambiguity, and domain-specific meaning. A frequently cited example is a word like bearing, which means something very different in engineering than in casual speech, or discharge, which in a medical setting could mean a patient leaving hospital or a wound leaking fluid. Human translators interpret the broader context and choose the correct sense, while raw machine output may pick a generic and sometimes dangerously wrong term. In legal and medical documents, such errors carry real risk.

Literary and creative work is even harder for machines. Different human translations of the same novel vary in tone and pacing because translation is interpretive, not mechanical. When a Dutch publishing house announced plans to use AI for translating commercial fiction, it triggered strong backlash from authors and translators who argued that machines cannot reproduce the charm and emotional resonance of skilled human work.

Pressure on translators and fair pay 🔗

The economic impact on professional translators is the sharpest challenge. Because machine translation is so cheap, clients increasingly expect lower prices, which squeezes the livelihoods of skilled translators. Industry tracking suggests translator rates have dropped significantly in many language pairs since 2023, and surveys of professional translators report that many have already lost work to generative AI. Balancing affordable automation with fair compensation for human expertise has become an ethical and practical problem the industry has not yet resolved.

The low-resource language gap 🔗

Machine translation also performs unevenly across languages. It works well for well-resourced pairs like English and Spanish, where vast training data exists, but quality drops for low-resource languages that have less digital text available. This is directly relevant to a multilingual society where many regional and minority languages lack the large datasets that modern systems depend on. Finding qualified human translators for these languages is difficult, and machines do not yet fill the gap reliably.

The hybrid future 🔗

The emerging consensus is not replacement but collaboration. Many organisations now use a hybrid approach, where machine translation produces a first draft and human translators refine it through a process called post-editing. This combines the speed of automation with the judgement of expertise. The future of the field appears to lie in this partnership rather than in either humans or machines working alone. Notably, official labour projections still anticipate continued demand for translators and interpreters, suggesting the profession is changing shape rather than vanishing.

What do you think? If machine translation keeps improving, should institutions invest in training more human translators for low-resource regional languages, or focus on building better datasets to feed the machines? And given the decline of centralised translation pools, is a global bibliography like the Index Translationum still the best way to track the world’s translations, or does it need a fundamentally new model for the digital age?

References

1. https://www.fbi.gov/how-we-investigate/intelligence/national-virtual-translation-center
2. https://www.crisoltranslations.com/our-blog/will-machine-translation-replace-humans/
3. https://www.unesco.org/xtrans/
4. https://library.mcmaster.ca/databases/index-translationum-international-bibliography-translations
5. https://datalinks.fandom.com/wiki/Index_Translationum_-_Unesco_Bibliography_of_Translations
6. https://www.ntm.org.in/
7. https://en.wikipedia.org/wiki/Central_Institute_of_Indian_Languages
8. https://gtelocalize.com/can-machine-translation-replace-human-translation/
9. https://latinobridge.com/blog/why-ai-cant-replace-human-translation-services/