Last week, an AI summit devoted considerable attention to language. We also marked International Mother Language Day. And this week, with the surrender of a senior Maoist leader in central India, a long and violent chapter in the region’s history appeared to reach an informal closure.

You may ask: what do large language models (LLMs) have to do with Maoism?

At the AI summit, one of the central discussions revolved around how vulnerable languages can benefit from advances in AI. Ironically, the summit itself reflected the global politics of language. The French are known for responding in French even when addressed in English. Yet this time, the French President delivered his speech in English.

India’s Prime Minister spoke in Hindi. Brazil’s President spoke in Portuguese. “Sovereignty” was a recurring theme throughout the summit. A journalist travelling with the French delegation remarked to me that since AI largely operates in English, President Macron chose English for strategic reasons. He then asked, “What exactly is ‘Bharat’?”

I explained that even our country’s name underwent Anglicisation, and there are now slow efforts to restore its indigenous usage as “Bharat.”

The Indian Prime Minister demonstrated how AI-driven translation tools could allow everyone to speak in their mother tongue while machines handle real-time translation. For major languages like English and Hindi, this promise already appears tangible.

But for smaller languages like Gondi, the road ahead is far longer.

The language of a conflict

For nearly two decades, Gondi was the primary language of the Maoist movement in central India. Around 500,000 Adivasis were associated with the conflict in some form. Approximately 5,000 people lost their lives. Over 17,000 were jailed. Nearly 20,000 eventually surrendered.

The Maoists’ symbolic capital was Abujhmad — a dense, forested, hilly region whose name literally suggests “the unknown.” It was not just geographically inaccessible; it was linguistically inaccessible. Gondi, the dominant local language, remained unintelligible to mainstream India.

This week, Chhattisgarh’s finance minister presented the state budget, announcing plans to build an “Education City” in regions like Abujhmad. However, when reminded that funds allocated two years ago for Gondi-medium education in schools had still not been released, he responded that he would “make every effort.”

The stakes are high. Bastar’s school dropout rate is nearly double the state average. For years, Maoist leaders told villagers: “We speak Gondi. The teachers speak Hindi.”

Gondi belongs to the Dravidian language family. Many Gondi-speaking children enter school without understanding a single word of Hindi. Unsurprisingly, many drop out early. At its core, the Maoist movement in Bastar drew heavily from these school dropouts, even though its leadership was often external and its ideological goals extended far beyond local linguistic grievances.

AI and the future of Gondi

At the AI summit, experts noted that work has begun on building translation models for smaller languages like Gondi. The government has launched applications such as Adivaani to support tribal languages.

However, language activists argue that the state is attempting to standardise only one dialect of Gondi. For years, scholars and community leaders have worked towards developing a composite, inclusive standard that reflects its multiple dialect clusters spread across six Indian states.

Today, three out of four Gonds no longer speak Gondi fluently. Yet comparisons are frequently drawn with Hebrew. During his visit to Israel this week, the Prime Minister addressed the Knesset — Israel’s parliament — where Hebrew is the principal language. Hebrew had remained largely dormant as a spoken language for nearly 2,000 years before being revived as a modern national language, supported by state policy and educational reform.

Some scholars argue — though definitive linguistic dating remains debated — that Gondi may be nearly 5,000 years old. Other Dravidian languages, such as Tamil, flourished because they received sustained state patronage. Gondi-speaking populations, by contrast, were divided across Madhya Pradesh, Chhattisgarh, Maharashtra, Telangana, Andhra Pradesh and Odisha. No single state assumed responsibility for its systematic development.

Madhya Pradesh and Chhattisgarh together have the largest Gondi-speaking populations.

This week, surrendered Maoist leader Venugopal stated in an interview: “It is a misconception that Adivasis joined Maoism because of jal, jungle, zameen (water, forest, land). They joined to fight injustice.”

Among those injustices was linguistic exclusion.

Beyond security: A linguistic solution

If the Union Home Ministry now seeks a durable peace in central India, addressing historical linguistic marginalisation should form part of the strategy. AI offers an opportunity.

The central government has initiated platforms like Adivaani. But such efforts must meaningfully involve the Gondi community and adopt a composite standard language framework rather than imposing a narrow dialectal version.

Meanwhile, Chhattisgarh’s proposal to build Education Cities in regions like Abujhmad must align with the National Education Policy’s commitment to mother-tongue instruction in early schooling. Ensuring that Gondi-speaking children receive primary education in their own language could drastically reduce dropout rates.

AI can then serve as a bridge — translating Gondi into Hindi, English and other global languages, and vice versa. Rather than replacing local languages, AI can help them connect to the world.

Lessons from history

International Mother Language Day, observed on 21 February, commemorates the five students killed in 1952 while protesting Pakistan’s attempt to impose Urdu over Bengali. That movement later evolved into Bengali nationalism, culminating in the birth of Bangladesh in 1971.

In a sovereign, modern democracy, no Gondi linguistic movement should need to follow such a path.

The question is whether India can use the tools of artificial intelligence not merely to digitise dominant languages, but to democratise linguistic opportunity — before silence becomes permanent.

END OF ARTICLE