26-04-2025
Opinion We can't just be users of AI. We have to be its co-creators
Also by Pramath Raj Sinha
The world is watching India. We are home to 16 per cent of the world's AI talent, with the fastest-growing developer population globally. Our AI skill penetration ranks ahead of the US and Germany. Homegrown models such as Sarvam-1 and Hanooman are making headlines. The IndiaAI Mission promises a vast expansion of infrastructure, research, and access. But amid all this promise, we must ask: AI for whom? For many Indians — shopkeepers in Cuttack, farm workers in Vidarbha, nurses in Dharwad, gig workers in Gorakhpur — AI remains distant, and at times, unusable. Earlier this year, one of us met Shabnam, a community health worker from Mumbai. Curious about AI, she tried using a chatbot to help her translate elaborate medical explanations to simple, everyday language that she could use to educate her patients. But the bot failed to understand her dialect. Her queries returned gibberish. Frustrated, she moved on.
The failure wasn't hers. It was ours. A tool that promises simplicity offered her complexity. India's AI journey is bifurcating. In one India, generative AI is a productivity multiplier. In another, it's unintelligible noise. We must not mistake availability for access. Building models and pushing tools into the public domain is not the same as building capacity. For AI to become meaningful across India's social fabric, we need a radical shift in our imagination of AI literacy – not as a technical curriculum, not as a skilling programme, but as a democratic right. What does it mean for a vegetable vendor in Surat to know whether the 'AI-based' pricing app he downloaded is making fair suggestions? Or for a woman leading a self-help group in Jharkhand to understand how AI might shape her creditworthiness?
This is not about teaching machine learning. It's about building critical agency so that citizens can question, interpret, and make informed decisions when AI touches their lives. Right now, most can't. The apps don't speak their language. The interfaces assume digital fluency. The data powering recommendations is blind to the local context. These tools, promising empowerment, often reinforce exclusion. Fixing this isn't about translating apps into more languages. It's about redefining development models — from top-down deployment to bottom-up co-creation. Effective AI literacy must be situated — rooted in context, community, and everyday needs. That means building hyper-local AI tools through participatory design. It means training 'AI ambassadors' in constituencies who can act as bridges between communities and technology. It means creating feedback loops where users shape how systems evolve. We have seen this model work before. India's digital financial revolution succeeded not because of an app, but because of the hosts of facilitators — bank sakhis, BC agents, WhatsApp groups — that made new tools intelligible, trustworthy, and usable. Why should AI be different?
Meaningful initiatives are already underway. Sarvam-1 supports 11 Indian languages. Hanooman supports 12. Microsoft has committed to training 10 million Indians in cloud and AI skills by 2030. And language learning is just the beginning. We need to capture local customs, accents, workflows, and knowledge systems by investing in datasets that reflect India's pluralism — not only scraping pages on the internet, but also collaborating with rural and urban teachers, community leaders, non-profit volunteers, municipal staff, corporate professionals and many other stakeholders from across the spectrum to co-create content that AI can learn from.
Let us be clear: There is no such thing as a neutral algorithm. And there is no such thing as an inclusive AI system without inclusive design, inclusive data, and inclusive dissemination. The AI literacy we need is not about knowing how ChatGPT works. It is about knowing when it does not. It is about helping citizens ask: Who built this? For whom? On what data? With what consequences? We cannot afford to wait for that flawless model. What we need is a movement, grounded in community, guided by trust, and fuelled by feedback. If India gets this right, we won't just be users of AI. We'll be its co-creators.
