Key Facts

An article traced how Tamil computing grew 25 years ago through campaigns led by academics such as former Anna University Vice-Chancellor M. Anandakrishnan, and noted that AI now needs billions of tokens of rights-cleared Tamil text and about 5,000 hours of annotated speech capturing its dialects. Tamil was declared India's first classical language in 2004 and has over 2,000 years of literary tradition, yet most commercial AI systems treat it as a low-resource language for want of large, high-quality datasets. The Tamil Valarchi Kazhagam runs the 'Sol Puthithu' programme to coin Tamil words for new terms, and the Tamil Virtual Academy holds digital Tamil texts.

GK Context

A 'classical language' designation by the Government of India (criteria set in 2004, later revised) requires high antiquity of early texts (1,500-2,000 years), a body of ancient literature considered valuable heritage, and literary tradition original to the language. Tamil (2004) was the first language so designated, followed by Sanskrit, Telugu, Kannada, Malayalam, Odia and others. In AI/NLP, a 'low-resource language' is one lacking sufficient digitised, annotated text and speech data for training large language models, which affects the quality of AI tools available in that language despite its cultural richness.