June 27, 2026, (Inside AI) — Maharashtra is deploying artificial intelligence to decode the origins of thousands of prehistoric petroglyphs etched across the Konkan coast, a project that could propel the sites toward permanent UNESCO World Heritage recognition. The state has earmarked Rs 15 crore for a four-year digital assessment led by IIT Madras’s technology hub, IITM Pravartak.
These rock carvings—depicting turtles, elephants, peacocks, and rhinoceroses—span 150 square kilometres across 107 villages in Ratnagiri and Sindhudurg districts. Archaeologists date them between 20,000 BC and 10,000 BC, older than the Harappan civilization. The carvings average 50 square metres by 20 square metres in size.
The petroglyphs were added to India’s tentative UNESCO list in 2025, but permanent status demands rigorous documentation. Tejas Garge, director of Maharashtra’s Directorate of Archaeology and Museums, told The Indian Express:
“These petroglyphs are the earliest evidence of art in the Indian context on such a large canvas. The figurines clearly depict the cognitive ability of human beings to visualise nature and replicate them in the form of these carvings.”
Read: AI Unwraps 2,000-Year-Old Herculaneum Scroll, Revealing Lost Stoic Text
The absence of horse and bull figures anchors the timeline. Horses arrived in the subcontinent around 1,500–1,000 BC, and agriculture began near 9,000 BC. Garge explained:
“The absence of horse figurines indicated that the carvings were created during the pre-historic era and the absence of bulls established that they were created during the pre-agricultural era. Therefore, it became clear to us that the origin of these carvings dates back to the Mesolithic era or the middle stone age period.”
AI Steps In Where Carbon Dating Fails
Traditional carbon dating is unreliable here. Coastal erosion and geomagnetic properties of the rocks skew results. Shankar Raman, CEO of IITM Pravartak, said the project will use drones for high-resolution aerial photography, then feed images into an AI system that identifies patterns, fills eroded lines, and compares motifs with global petroglyphs from civilizations like Mesopotamia.
Raman told The Indian Express:
“The petroglyphs are spread across coastal Maharashtra that is also known for receiving very heavy rainfall. Therefore, there are portions of stone carvings that are already eroded. The rocks also have geomagnetic properties, which pose a problem in carbon dating.”
The AI will cross-reference carvings with local tribal traditions, folklore, and food habits. Volunteers are already interviewing communities. The study may also reveal the region’s ancient flora and fauna.
Local resident Sudhir Risbud, who first spotted the carvings in 2007, has documented over 3,000 figures through his nonprofit Nisarga Yatri. He said manual documentation is impossible given their scale, and high-end drone cameras are essential.
Read: Global AI models struggle with Indian wildlife sounds. 59 volunteers built a fix.
The UNESCO Dossier and a 2030 Deadline
To secure UNESCO status, India must submit a detailed dossier covering historical significance, conservation plans, and global comparisons. The Archaeological Survey of India’s nomination slots are booked until 2028–29, so Maharashtra aims for a 2030 pitch.
Garge said the IIT Madras study will compile all required data. The project’s timeline aligns with that goal. If successful, the petroglyphs would join India’s 43 World Heritage sites, reinforcing the Konkan coast as a cradle of prehistoric art.
The initiative also tests AI’s role in archaeology beyond pattern recognition—linking art to cultural memory and environmental history. While some experts caution that algorithmic comparisons risk oversimplifying cultural contexts, the project’s interdisciplinary approach could set a precedent for heritage documentation in tropical, erosion-prone regions.
Sofia spent years studying how machines parse human language before pivoting to cover the systems now rewriting it. She reads technical papers the way others read news and looking for what the numbers are actually saying beneath the headlines.
