Launching the AI-based Large Language Model (LLM), ‘Sanskrit Heritage Model’, at the Madras Sanskrit College here, Sitharaman said that language learning should draw people towards great literature rather than confine them to rigid grammatical rules.

“The language has to be tempting enough for people to learn. These are the kind of spice you need into it,” she said, cautioning against tools that are dry as bone in teaching Sanskrit. “We need for wider critical and appreciative studies of Indian classics, contrasting the extensive scholarship available on William Shakespeare with the relative scarcity of accessible analytical works on Sanskrit poet Kalidasa.”

She added that AI could help bridge this gap by analysing classical texts extensively. “A tool capable of explaining grammar, Sandhi and Samasa and identifying literary devices such as Upamas could provide scholars with an invaluable research aid,” she said.

Highlighting Sanskrit’s historical influence beyond India, she referred to the presence of Sanskrit words in contemporary Croatian and Indic-language repositories in the Philippines, Vietnam’s Cham kingdom, Central Asia and the Caucasus. She urged technologists to use AI to locate and index manuscripts, palm-leaf records and inscriptions scattered across global archives, citing the Leiden copper plates that documented grants to a Buddhist vihara near Nagapattinam.