The intention is to democratise use of the internet in India by allowing people access in regional languages. Officially, India recognises 22 languages with 12 scripts. Clearly, translation in scale for such a diverse set of languages to take internet content to the people is only possible through the application of deep tech - machine translation.
At present, access to the internet is mostly in English, though only 10% Indians are proficient in it. While there are a few startups catering to regional language preferences and some browsers that offer translations access to content on the internet for a non-English user is restricted.
Like all other digital initiatives undertaken by India, this, too, is based on an open digital architecture. As a result, both in scale and scope, the latest digital initiative is the most ambitious ever. It is like a UPI (United Payments Interface) moment for digital inclusion.
For more than a decade after the use of the internet took off, English was the primary language of access. However, by the turn of the millennium, access to the internet had begun to be enabled in other international languages, gradually eroding the hegemony of English. According to Internet World Stats, the two most prevalent languages on the internet are English (25.9%) and Chinese (19.4%). Spanish and Arabic are a distant third and fourth at 7.9% and 5.2%, respectively. No Indian language made it to the top 10.
English Vinglish
The political economy of this is obvious: proficiency in English determines the scale of access to the internet in India, further worsening the existing digital divide. Enabling access in regional languages will, therefore, democratise internet use in the country. In fact, a GoI white paper on Bhashini (bit.ly/3BdMuCg) reveals that more than one in two of those surveyed said they would use the internet if the content was made available in local languages.
The rollout of Bhashini was formally proposed by Nirmala Sitharaman in her 2021-22 budget. 'We will undertake a new initiative - National Language Translation Mission (NTLM). This will enable the wealth of governance- and policy-related knowledge on the internet being made available in major Indian languages,' she told Lok Sabha. Since then, NTLM has acquired the moniker Bhashini and was launched this year on the seventh anniversary of the Digital India week.
To be sure, the idea of deploying machine translation to translate content from English to other languages has been in the making for decades. Though India came late to the party - while the western world launched this effort in the 1950s - researchers at the Indian Institute of Technology (IIT) and the Tata Institute for Fundamental Research (TIFR) had started exploring it from the 1980s with reasonable success. It got wings when the earlier avatar of the ministry of electronics and information technology (MeitY) started funding R&D projects and set up the Technology Development for Indian Languages (TDIL) in 1991.
Making Tongues Wag
After 2005, there was a fortuitous convergence of several trends, which provided an unexpected boost to this initiative:
n The advent of neural processing in which computers acquired the ability to refine their output tremendously by being able to process more information using artificial intelligence.
n The launch of smartphones and their proliferation empowered users. In the post-Jio world this meant easy and cheap access to data.
n Broadband connectivity under the National Optical Fibre Network (NOFN) is now extending its footprint to rural India. As on July 1, 181,216 of the 262,825 gram panchayats in the country are now part of the optical fibre grid.
Throughout most of this period - when the ecosystem was being developed - the research was driven by the TDIL. In fact, there have been several notable successes. The most high profile is the Supreme Court Vidhik Anuvaad Software (Suvas). It translates the judgments delivered in English into nine major Indian languages and vice-versa. It has since been adopted by the Bangladesh judiciary, too.
In addition, there exist commercial translation systems offered by Google, Microsoft, Amazon and Facebook. Though remarkable, it is nowhere close to the desired levels of scale in universalising content in regional languages. With the launch of Bhashini, NLTM has gone into mission mode to resolve this asymmetry of digital access in India.
No comments:
Post a Comment