Translation tools end language barriers

Advanced new technologies are bridging the gap between different languages: It could be by assisting train passengers in India or managing policy documents in Europe.

While language translation software has been available since the 1950s, it was mostly used by researchers and professionals. When Google Translate was launched in 2006, instant language translation became available to millions. The world was taken by storm. 

The second phase of the translation revolution is being led by the rising popularity of generative artificial intelligence (AI) tools like ChatGPT, the popular chatbot that responds to a range of written queries.  

Many professionals and institutions are increasingly using such tools for translation. Entire documents can be translated in minutes with few errors. These can be policy documents, court judgments, medical research or content that needs specialised attention. 

While Google Translate does literal translations, GenAI tools can also add context and improve grammar. These chatbots work on an iterative concept and engage with the user. Consumers can make follow-up requests and provide feedback. Different versions of a translation can also be done for a user. 

European Union (EU) institutions have 24 official languages. The bloc’s translation department is unique for its ability to ensure that all documents are available in every official language. Recently, the EU has begun to use new technology tools to improve and speed up quality translation. The workload of translators had increased to 2.5 million pages in 2022, according to estimates. Now, AI-based tools will translate press releases and other such documents in an automated format for public consumption. EU is using a system that deploys artificial neural network (an AI method that teaches computers to process data in a way that is inspired by the human brain) to understand and predict worlds in a relevant sequence. 

The global language translation software market was valued at $9.3 billion in 2021 and it is projected to reach $44.8 billion by 2031, according to a report by Allied Market Research. “Various companies and government bodies are collaborating to strengthen R&D [research and development] in the many industries with evolving productive alliances that lead to indigenous design, development, manufacturing, and deployment of cost-effective products and solutions. Additionally, government organisations also adopt language translation software for different purposes such as immigration services, diplomacy, and foreign trade,” it said. 

Phrase, a translation software firm, is applying new technologies for its services. It has announced “a range of enhancements to the Phrase Localization Suite that will define the next generation of localization technology, powered by AI.” 

Phrase will deploy machine learning (ML) models and customised machine translation (MT) engines that are optimised to meet specific business needs. This will help its clients achieve better translation quality and speed. The new tools can also be customised to reflect the terminology, style, and tone of voice for a specific individual or entity.

India’s Bhashini project brings various languages together on its translation platform. This is an outcome of the National Language Translation Mission launched a couple of years ago. AskDISHA, a chatbot launched by the Indian Railway Catering and Tourism Corporation (IRCTC), is built using Bhashini. It can answer questions in English, Hindi and Gujarati while 10 languages are being added. Consumers will be able to connect, query and transact using this bot in their own languages.

As MT and ML tools become more sophisticated, they will be able to offer better and localised translation. From languages, they will be able to move on to dialects and provide social context to translations. Such capabilities can demolish the barriers between content and knowledge in various languages. 

A new layer of voice and listening tools can further add to the translation revolution powered by AI.