Heritage tech: AI helps keep traditional knowledge alive and accessible

From BharatiyaGPT to tribal language apps, AI is reshaping how India preserves, decodes and revitalises its ancient knowledge systems

Artificial intelligence (AI) is transforming the preservation of ancient knowledge, from digitising fragile artefacts and deciphering endangered languages to enabling vivid virtual reconstructions. AI tools are being engineered to capture and safeguard Indian heritage. 

Developed by ImmverseAI’s BharatiyaAI Foundry Lab, BharatiyaGPT is a comprehensive AI ecosystem designed as India’s answer to global models. Its core mission is to digitise the Bharatiya Gyan Parampara (Indian knowledge systems) while correcting the inherent biases found in Western-trained AI. 

The platform is built on a large language model trained on eight billion parameters to understand, generate and predict human-like text and code. Its strength lies in its expansive, culture-focused data pipeline that is digitising and processing over 50 terabytes of ancient Indic manuscripts. 

The effort encompasses more than a million texts across 23 languages and aims to decode 64 ancient scripts, including Devnagari, Sharada and Grantha. The process is set to unlock over one trillion unique Indic tokens, establishing an authentic, context-aware foundation for the AI. BharatiyaGPT also supports 100 world languages. 

Its capacity ensures that cultural concepts are conveyed accurately and with linguistic nuance. The platform has specialised AI models like AyurvedaGPT for traditional healthcare, GanitGPT for mathematics, and SthapatyaGPT for architecture. A key feature of BharatiyaGPT is its Revolutionary Learning Ecosystem, which includes the World's First AI-Infused “guided learning book”. The book goes beyond traditional e-reading to offer interactive content and AI-powered mentorship, transforming passive study into an active, conversational journey through ancient texts like the Vedas and Upanishads. 

GanitGPT decodes concepts from Vedic mathematics, geometry, and computations, while AyurvedaGPT extracts actionable insights on doshas (ailment), remedies and wellness from traditional texts. Key features of the second tool include personalised health guidance, support for 23 Indian languages, and recommendations on treatments and herbs derived from over one million manuscripts. “Over 25 scholars are working alongside 100+ AI/ML [machine learning] engineers to unlock India’s manuscript heritage,” said R Ramakrishnan, cofounder and chief executive officer of ImmverseAI. 

Adi Vani, an AI-powered app launched by the Ministry of Tribal Affairs in 2025, documents, translates, and revitalises endangered tribal languages. It provides real-time text and speech translation between Hindi, English, and languages like Bhili, Gondi, Santali and Mundari, with plans to include Kui and Garo.

 

The app’s interactive modules teach foundational language skills, while digitising folklore and oral traditions creates searchable archives. Developed by the Indian Institute of Technology Delhi in consortium with tribal research institutes and other engineering colleges in Jharkhand and Odisha, Adi Vani involved more than 250 native speakers translating NCERT books and local stories. Community feedback continuously refines the AI at low cost.

 

By subtitling government advisories, health messages and speeches in tribal languages, it bridges communication gaps and boosts education in remote areas. It helps preserve 82 “vulnerable and 42 critically endangered languages” of India’s 104 million tribal people.

 

Beyond India, AI is driving global heritage protection for public consumption. Algorithms like Ithaca restore damaged ancient Greek inscriptions with 60 per cent accuracy — rising to 70 per cent with human collaboration — following training on 60,000 texts spanning 700 BC to AD 500. In Italy, the RePAIR project employs AI and robotics to reassemble thousands of Pompeii fresco fragments by analysing shapes, sizes and illustrations.

 

Machine learning deciphers eroded scripts, cuneiform, hieroglyphics, and the Dead Sea Scrolls by spotting patterns and identifying individual scribal hands. Optical character recognition and natural language processing transcribe faded manuscripts into searchable digital text, while high-resolution 3D scanning and predictive modelling detect structural erosion in monuments for proactive conservation. Furthermore, AI dates artefacts through chemical and stylistic analysis, identifies potential excavation sites via geographical data, and revitalises endangered languages by archiving oral histories through speech recognition.

 

Tradition and technology are coming together constructively for an informed future.