Status of AI in Indian Languages

India, a multilingual country has 1369 mother tongues including 22 scheduled languages (as per 2011 Census reports). Language aided applications based on AI technology are mainly developed based on the English language. Several initiatives have been and are underway to develop natural language processing (NLP) models, speech recognition systems, and other AI technologies that cater specifically to Indian languages. Efforts were being made to build datasets, create language models, and develop applications that can understand and generate content in languages such as Hindi, Bengali, Tamil Malayalam, Kannada, and others. However, challenges persisted, including the lack of extensive datasets, linguistic variations across different regions, and the need for specialized expertise in Indian languages within the AI research community.

Government and corporate-developed tools provide translations in major Indian languages. Translation quality can vary based on different factors. Beyond translation, services like language identification, speech-to-text, and text-to-speech, Chatbots are being developed. These features are more advanced in some languages, and major Indian languages are catching up. Efforts are underway to expand them to even minor languages and dialects.