Multi-Lingual Internet – For the Regional Language Users

The multi-lingual Internet will bridge the Digital Divide

Internet content in Indian languages for the non-English speaking majority will define the next phase of growth in India and other regions.

A lot of work has already been done in India towards the development and standardization of content and translation. Internationalised Domain Names (IDN) in all major Indian languages for websites and email. Several startups are solving problems in this area.

A panel consisting of India’s best brains in Language technologies and IDN got together to discuss the achievements and successes so far and the long road ahead.

Mahesh Kulkarni, formerly Senior Director, Corporate R&D, CDAC, and currently on a mission to promote the adoption of Artificial Intelligence (AI) and Machine Learning (ML) in language technologies has been tirelessly working in this field for nearly 3 decades.

Akshat Joshi, is the founder of ThinkTrans which is working in the area of Indian languages. Akshat has done some pioneering work in Indian Language computing — code conversion, font engineering and machine translation during his 11 year stint at CDAC.

Nakul Kundra, is the Founder of Devnagri, which offers a translation platform for all Indian languages.

The panel was moderated by Gajendra Upadhyay, Editor, Voice&Data.

What are some of the major initiatives by the Government in the Multi-Lingual Internet space.

The Integrated Devnagri terminal was a major step forward according to Mr Kulkarni. “Encoding for Indian languages, making keyboards available with Indian fonts - were foundational.”

Standards and keyboards for Indian languages gradually evolved. Slowly, a lot of companies started realizing the importance of Indian languages and started working in this direction. The Technology Development of Indian Language (TDIL) Programme of the Government is another example – it was set up with a mission to promote awareness of Indian languages. Many other projects with the best institutes of India bolstered this focus. The government has played an important role with many initiatives for standardization and compliance.

Akshat felt that adding to the depth was the availability of apt content, Internet-based services in Indian languages, democratizing Internet access points with internationalized domain names and emails. But there is still a gap.

One of them is the issue of access of these websites due to the technical discrepancy in top-level domains. “It is a chicken-and-egg problem because most people have to use an email ID as a gateway to creating a profile on these sites. But there is a barrier in the acceptance of language email with IDNs. The universal acceptance (UA) initiative of the ICANN will help to address the gap.

“Ninety percent of Indians,” said Nakul, “are not fluent in English.”

But content availability in Indian languages is abysmally low. Nakul has created a translation engine with contextual capabilities. “We help translate content for domain-specific needs - this area is a huge opportunity.” There are many government mandates for every industry and there is a push to go vernacular now.

Mr. Kulkarni also emphasized that data cannot be seen in isolation as we see the impact of AI, ML, and new paradigms like Metaverse. “Voice-driven applications would be on the top, backed by ML transformation. Accessing the Internet will help people with not just entertainment but services, education, and healthcare. A good example is that of Voice Assistants in banking. These systems can understand the intent of customers in a good way now.”

Pain areas in multi-lingual transformation are technology enablement and policy aspects, in Akshat’s view. Services and people-interaction are driven by policies at various layers of the organization.

“To read an email from a different language, understanding of script needs both a technology and policy-level solution.”

Nakul also had a view on this. Colloquial translation gaps create a need for catering to use-cases beyond mobile apps, real-time voice translations, ticketing services, etc. “Once this ecosystem is created for Indian languages, businesses can think of this area in a different way.”

Mr Kulkarni drew attention to some policy areas that need to be mandated. For e-governance platforms, availability of language keyboards, various fonts that reduce data loss, universal acceptance, three-language formula for government websites etc.

Developing the three-language formula for mobiles is required. The collection of data and solutions through crowd-sourced models will capture the nuances in a comprehensive manner. “These issues, with constant work, and over a period of time, will be resolved.”

Nakul felt that this is how we can bring the next billion on the Internet.