No One Can Predict the Jobs of 2035. So What Should We Be Teaching?

This is Part 3 of a 3-part series on India's demographic dividend in the age of AI. Part 1 laid out the credential crisis. Part 2 reflected on inability of the richest's countries to automate their way out. Today: what actually needs to change, and it starts before the boardroom.

There are 260 million children currently in Indian schools. Every single one of their parents shares the same unspoken fear: what if everything we're teaching them becomes irrelevant?

It's not an irrational fear. By 2030, 70% of job skills are expected to change. The WEF estimates 92 million jobs will be displaced globally. India's own Economic Survey 2025-26 acknowledged that 35-40% of current jobs are prone to AI-powered automation. A third-grader starting school today will graduate into a world where AI is as fundamental as electricity, and we have no reliable map of what that world's job market looks like.

For most Indian parents, the purpose of education is singular: job security. Get a degree. Get a job. Build a life. But what happens when no one can predict what jobs will exist in a decade?

The Uncomfortable Truth About Skills

This week at the AI Impact Summit, $200 billion in AI investments were announced. Anthropic, Google, Microsoft, Adani, Blackstone, everyone is pouring capital into India's AI infrastructure. But as Vinod Khosla warned from the same stage: IT and BPO services could disappear within five years. The roles that gave millions of Indian families their first taste of middle-class security are the exact roles AI replaces first.

So we can't prepare children for specific jobs. We can't even prepare them for specific industries. What we can prepare them for are the capabilities that remain valuable regardless of what AI does next.

And here's the counterintuitive insight from this entire series: those capabilities aren't technical. They're deeply, irreducibly human.

What Should We Actually Be Teaching?

India's Economic Survey itself pointed to the answer. It identified leadership, entrepreneurship, teaching, and caregiving as fundamentally AI-resistant; not because they're low-skill, but because they depend on moral judgement, trust, and cultural nuance that software cannot replicate.

MIT Sloan's EPOCH framework puts it more precisely: Empathy. Presence. Opinion/Judgement. Creativity. Hope. These five capabilities are what AI struggles with most, and what the world's labour markets will reward most.

Now look at what we're actually teaching. 93% of Indian students know only 7 career paths. Only 7% receive formal career guidance. 65% pursue degrees misaligned with their own aptitude. We're optimising for an industrial economy that no longer exists, using an education system designed before the internet, let alone AI.

The shift needed is fundamental: from training for jobs to developing for capability.

Teach a child to code, and you give them a skill with a half-life of five years. Teach a child to think critically, communicate persuasively, collaborate across differences, adapt to ambiguity, and exercise ethical judgement, and you give them a toolkit for any future.

These aren't "soft skills." That label has always been a disservice. They're the hardest skills to develop, the hardest to automate, and the most durable across every economic cycle in history. Every parent who worries about their child's future should know: the single best investment isn't a specific degree. It's these foundational human capabilities.

The Convergence

Here's where the three parts of this series come together:

Part 1: India has 800 million young people, but it's producing credentials, not capabilities. The education-to-employment pipeline is broken.

Part 2: The world's richest economies are running out of people and discovering AI can't replace them. They need humans who can work alongside AI, and India has the world's largest supply.

Part 3: The bridge between India's people and the world's demand is capability development at scale. Not more degrees. Not more coding bootcamps. The foundational human skills that no technology cycle has ever made obsolete.

The demographic dividend was never about headcount. It was always about what's in the heads. India's 800 million aren't an asset because they exist. They become an asset when they can think, adapt, lead, and do what machines cannot.

The clock is running. The window stays open until 2045. And the answer was never GPUs.

It was always people.

Sources: WEF Future of Jobs 2025, India Economic Survey 2025-26, MIT Sloan EPOCH Framework 2025, AI Impact Summit 2026 (New Delhi), OECD Employment Outlook 2025, Mercer-Mettl 2025.