India turns to AI entrepreneurs to decode Harappan script

For over a century, the Harappan script – one of the world’s oldest and most mysterious writing systems – has defied scholars’ efforts to decipher it. Found on thousands of seals and artifacts across ancient sites such as Harappa and Mohenjo-daro, the script’s meaning remains unknown despite decades of studies.

In addition to manual methods, India is now looking to young entrepreneurs in the fields of AI and machine learning to help decode the Harappan script.

The agenda would be part of discussions at the formal launch of "Gyan Bharatam Mission" that aims to preserve the country's rich manuscript heritage. The launch would happen at an international conference planned here from September 11 to 13 by the Ministry of Culture and Indira Gandhi National Centre of Arts. Titled, “Reclaiming India’s Knowledge Legacy through Manuscript Heritage”, the conference will bring together historians, conservationists, and technology experts to explore how manuscripts and inscriptions, including the Harappan script, can be preserved, digitised and decoded.

Experts say several factors make decipherment of the Harappan script particularly difficult. The inscriptions are often extremely short – some containing fewer than five signs – providing little context for interpreting language or grammar. Unlike Egyptian hieroglyphs, there is no bilingual text to assist researchers in cross-referencing the script with a known language.

To share findings of their attempts to decipher the Harappan script, more than 1,100 participants from all over India are attending the conference. Among them are over 95 academics, 22 government and private sector officials, 179 professionals, 112 research scholars, 230 students, and more than 400 individuals from diverse backgrounds.

"After a pan-India competition was organised to  identify young entrepreneurs who are working in the area of AI and machine learning and who could be instrumental in decoding & deciphering the manuscripts, only 40 participants and 10 researchers to take part in it (in the conference)," said the Ministry of culture on Wednesday.

However, AI-based efforts have already been underway. A recent March 2025 study using a hybrid CNN-Transformer model explored visual patterns in Harappan inscriptions and found symbol frequency and co-occurrence that align with known scripts, but the researchers concluded that further linguistic validation is required to draw definitive conclusions.

The language behind the script remains a mystery, with theories pointing to Dravidian, Brahmi or other lost dialects but there is no consensus in sight.

The script also appears exclusively on seals, pottery, and other durable materials, with no evidence of inscriptions on palm leaves or bark that might have preserved longer content. Interpretations vary widely, with some scholars suggesting it served administrative functions, while others believe it was primarily symbolic or religious. Early this year, MK Stalin, Chief minister of Tamil Nadu, announced a $1 million prize for anyone who can decode the Harappan script.

GYAN BHARATAM MISSION

The mission, which will be formally launched on September, has been allocated Rs 482.85 crore from 2024 to 2031 with the purpose of documenting India's 10 million manuscript texts. So far, over 44 lakh manuscripts have been documented and preserved.