AI everywhere, direction nowhere!

AI to the left of us, AI to the right of us, AI in front of us — and we haven’t yet stopped to ask where we’re steering. The rhythm is deliberate; it echoes Tennyson’s cadence not to dramatise the moment, but to remind us that charge without formation is how armies are lost, not how futures are built. Presence is not the same as direction.

The moment feels overwhelming. Every company now claims to be “AI-powered.” Every startup pitch begins and ends with a large language model. Every consumer device, from air-conditioners to pressure cookers, now advertises intelligence as though sprinkling spice.

It is easy to mistake ubiquity for progress. But this moment requires less celebration and more orientation. At Trafalgar, the sea was loud, the cannons louder, but victory came not from the most noise — it came from the navies that held formation. Momentum alone does not create history. Discipline does.

Right now, discipline is in short supply

AI has moved from breakthrough to fashion statement. The Prime Minister’s clever play of “AI can also mean American/Indian” landed because it made the technology feel personal, accessible, even patriotic. Once leadership signals that something is the future, the country does not walk — it rushes. Some rush to build. Many rush to re-label.

And so we now have AI-enabled job interviews where résumés are rejected automatically for a spelling quirk. “AI news anchors” reading scripts prepared manually minutes earlier. AI projectors in classrooms that are just autoplay settings under a different name. Even the modest microcontroller has been rebadged into an “AI chip inside.”

This is not transformation. This is spray-on futurism — the label spreading faster than capability.

Yet, beneath the theatrics, something meaningful is happening. AI is quietly making daily life less exhausting. It drafts the email you didn’t want to write. It summarises the document you did not have time to read. It translates a prescription for a caregiver who speaks another language.

It lets a student revise a concept slowly, rather than pretend to understand in a classroom of sixty. It helps a farmer identify blight with a single photograph instead of waiting days for an agricultural officer. AI is not replacing the human. It is removing friction. It gives us minutes back. And minutes matter.

The real danger is not that India will invest too little in AI. It is that we may invest too quickly, confusing speed for strategy.

This is where the latest policy announcement — introducing Artificial Intelligence and Computational Thinking in schools from Grade 3 — becomes instructive. The intent is forward-looking: equip every child with reasoning, pattern recognition, and problem-solving skills; move away from rote memorisation towards cognitive flexibility. This aligns with NEP 2020 and the global education shift.

But intent is not capacity

Many government schools still struggle with stable electricity. Teacher training capacity is uneven. Foundational literacy and numeracy missions are still catching up after pandemic disruptions.

If AI education is introduced without preparing teachers, materials, and classroom support, it risks becoming another subject to memorise, display in charts, and reproduce during inspection — without genuine understanding. AI may become handwriting practice by another name.

The Women’s Reservation Bill was passed with applause — and then its implementation was left to an open-ended future. Announcing AI education nationwide without preparing the ground risks the same fate: a grand promise celebrated today, and an uncertain, diluted reality tomorrow.

A wiser approach would be to run deep pilot programmes in a few states — rural, semi-urban, and urban — learn what actually works in real classrooms, refine, and scale. True reform grows outward from success, not downward from decree. Direction needs pacing. Speed without grounding is drift.

The deeper challenge is that our education, training, and employment systems were built for a world where knowledge lived in the mind. Exams were tests of recall — how much you could store and reproduce. But we are entering a world where information is retrievable instantly. Knowledge is no longer memory — it is navigation.

The real skill now is judgment: the ability to verify, interpret, differentiate signal from automation. If we do not shift education from memorisation to reasoning, AI will not democratise opportunity — it will concentrate it. The divide will not be between “tech jobs” and “non-tech jobs.” It will open within the middle class itself — between those who learn to use AI as leverage and those quietly outpaced by those who do.

Language deepens this divide. Most powerful AI systems still think in English. But Language deepens this divide. Most powerful AI systems still think in English. But India lives in its languages — from Bengaluru tech floors to Chhattisgarh haats, from Mumbai trains to Manipuri courtyards.

AI that cannot understand your tone, idiom, humor, grief, hesitation, or pride does not empower you. It works around you. The real AI race is not who deploys fastest. It is who builds models that understand their people.

Meanwhile, capital is flooding the space. Companies are buying GPUs because others are buying GPUs. Institutions are signing multi-year cloud contracts without securing data portability. Government departments are announcing “AI readiness” without securing maintenance budgets. This is how readiness turns quietly into strategic dependency.

The global AI ecosystem runs on closed loops. Microsoft invests in OpenAI; OpenAI runs on Microsoft Azure. Amazon backs Anthropic, which trains on AWS. Google provides TPU compute to the same labs it funds. NVIDIA invests in cloud providers that then buy NVIDIA hardware in volume.

These are not conspiracies — they are efficiencies. But when capital reinforces itself, prices begin to reflect belief rather than objective value. We have seen this movie before: telecom spectrum races, e-commerce land grabs, crypto booms. The technology was real. The stampede was the problem.

India now stands at the pivot where direction matters more than acceleration.

If we build our AI ecosystem purely by renting compute, platforms, and model intelligence, then our data will train someone else’s systems, our public institutions will pay subscription fees to foreign platforms, and our strategic choices will be shaped by pricing, not priorities. Our future becomes a license agreement.

But we can choose differently

We can build shared national compute clusters, accessible to universities and state boards.

We can compile responsibly governed Indian datasets, reflecting our languages and ways of reasoning.

We can train millions, not an elite, to use AI well — from nurses to clerks to drivers to teachers.

We can build AI systems that speak as we speak — not translated, but native.

This is not techno-nationalism. It is economic self-respect.

The hype has done its job. It brought imagination. It opened the room. Now comes the part that determines whether this moment becomes transformation or just another chapter of acceleration without architecture.

The future is moving either way. But we do not have to be passengers. We do not have to imitate. We can steer. The wheel is in our hands. The question is whether we choose to use it.

(The columnist is a Mumbai-based author and independent media veteran, running websites and a youtube channel known for his thought-provoking messaging)