An overview of automation in the translation ecosystem and how human input is more essential than ever
Translation can be a complex undertaking. It can involve any type of text in any style and can be destined for many different purposes. Translation – simply meaning the transfer of spoken or written communication from one language to another – can be a straightforward exchange of words or it can call for the reimagining of the content in a whole new cultural context. Translation is needed worldwide by businesses in all sectors and from all nations.
Faced with such diverse clients with such varied needs, language services providers (LSPs) have always been keen to adopt any tools to make their tasks easier. Technology has played a huge role in simplifying translation processes, particularly from the 1990s onwards when translation memories were first used commercially.
In recent years, the age of artificial intelligence (AI) has meant that technological change is happening more rapidly than ever before and the emphasis on reducing the time it takes to produce a high-quality translation has grown.
How does automation help the translation process?
Automation has been accelerating across the board in translation processes for a while and LSPs are investing significantly in different technologies. Machine translation (MT), boosted by AI and the arrival of large language models, is at present a clear focus and language companies are growing their emphasis on this and the role of human post-editors. Automating workflows is also currently a focal point for technology investment, as is the development of software connectors that facilitate exchanges between client and LSP systems.
Translation processes lend themselves to automation in key areas. Language itself can be automatically translated using machine translation, technology can automate unnecessary manual steps in translation workflows and connectors can make communications between LSPs and their clients run seamlessly and without the need for clunky file downloads or email exchanges.
Let’s look at a few of the tools that LSPs use at these different stages of the translation production line.
- Translation memory – these tools are a type of computer-aided translation (CAT) that store sentences or segments of translated text in databases. These segments can then be automatically reused if the same original text reappears, meaning no text is translated twice. Translation memory systems save translators considerable amounts of time that can then be passed on to the customer in cost efficiencies.
- Terminology management programs – keeping track of glossaries or term bases in Excel spreadsheets isn’t a viable option for any sizeable translation project and LSPs today use software programs to record and store industry- or company-specific words and phrases in all relevant languages. These terminology databases gather phrases, trademarks, product names, slogans, technical terms and so on, along with their definitions and notes on usage. They can be vital for maintaining consistent corporate language, especially across large organizations with many sites and employees.
- Connectors – these programs permit automatic exchanges between software systems, for example, between content and translation management platforms, enabling content to be translated automatically without having to manually collate, copy or upload it. Read about some of the dozens of integrations that t’works links to here.
- Machine translation and AI – for translation projects involving large volumes of text where time is of the essence, machine translation offers a good solution. Machine translation is particularly effective when domain-specific engines are optimized to individual project requirements with existing translation data. AI’s ability to constantly learn and improve means it is increasingly benefiting machine translation systems, which in turn are becoming more dependable. MT can be effectively employed for non-ambiguous, plain texts as found in, for example, e-commerce, social media and internal communications, and for more complex texts where a human post-editor is used.
- Translation management systems – these are sophisticated programs that offer the possibility of automating many stages of the translation process under one centralized system and include technologies to automate workflows as well as automatic translation.
- Customer portals – LSPs frequently provide their clients with customer portal interfaces to enable them to have complete visibility of all their interactions with the translation provider. Clients can then easily monitor and action inquiries, quotes, orders and deliveries in one place and benefit from encrypted exchange of large amounts of data. Learn about the t’works customer portal here.
Accuracy and consistency at all times
These tools improve the efficiency of the translation process by reducing the need for unnecessary human input and speeding up the time to delivery. Waiting weeks or even months for translations to arrive is a now thing of the past and means companies can translate the much higher volumes of content that they create today, more effectively and more swiftly. What’s more, as LSPs increase their productivity they are then able to pass on any savings in time and cost to their customers.
Translation accuracy is also improved. Tools like translation memory and terminology management are able to ensure that translations benefit from a sustained high level of precision, guaranteeing consistency in all languages across business texts at all times.
However, it’s important to note that these tools do not mean human involvement is no longer necessary. On the contrary, human expertise is more essential than ever and the growing use of automation frees up human experts to concentrate on the parts of the translation process where they are most needed.
The role of human proofreaders and quality control
For many translations, the first draft isn’t the final version and whether or not the translation is done by a machine or a person, proofreading it is always vital. Proofreaders spot any grammar or punctuation inaccuracies and any formatting and spelling errors. Translation proofreaders will be able to compare the original to the translated text and ensure accuracy and fidelity.
Post-editors perform a similar but more complex role to proofreaders and this is an expanding role within the language industry. With the rising quantity of machine translation output and the reality that MT engines are not able to understand all the subtleties of language in the same way native speakers can, post-editors carry out any further translation and revision to make sure the end translation is fit for purpose.
Other human roles are indispensable to the translation workflow. Sometimes the human touch is called for, particularly to iron out potential pain points and project managers are there when only person-to-person communication will do. Copywriters, native language linguists, editors and technology experts all contribute to the maintenance of quality across translation projects.
In-house human teams with experience and expert knowledge, together with the latest technology and optimized workflows make up the ingredients for successful translation processes. These essential ingredients will have more or less input depending on the type of translation and its intent.
Trends and developments
A few months ago many were predicting the demise of the human translator and the rise of machine-powered translation enabled by AI technology. As with other recent tech advances, generative AI has been overhyped in certain parts of the media and it’s only now, as the limitations of large language models like ChatGPT are more understood, that the hard work begins to transform this tech into practical, real-life solutions.
Translation will undoubtedly continue to develop in line with artificial intelligence (as it has been doing for many decades), but the emphasis will remain on finding the best balance between people and technology. The focus for translation services will be to make the process as glitch-free as possible for their customers, using technology tools to enhance and streamline operations for the benefit of translation buyers.
Human roles will continue to evolve and offer the expertise, artfulness and true language understanding that only people can provide. Human communication is at the heart of what translation is and it is people who will continue to develop, shape and deliver translation for the benefit of the people receiving it.
Look out for part 2 of our blog on translation automation which will take a deeper dive into some of these clever technologies and in particular how connectors simplify the process for translation buyers.
Talk to t’works today about how our human-technology workflows will help your business overcome language barriers and reach a new global audience.
