India’s AI education strategy is ambitious and unusually comprehensive, reflecting a seriousness of intent that is rare in large public systems. Yet, for all its scale it currently addresses only one side of the problem.

In just the past few weeks, India has launched an AI and computational thinking curriculum across more than 32,000 CBSE schools. Alongside this, national initiatives like SWAYAM, FutureSkills PRIME and the IndiaAI Mission signal a coordinated and large-scale push toward AI capability.

Taken together, this is definitely a serious national effort. But when one looks closely, a pattern becomes difficult to ignore. These initiatives are designed to introduce AI capability as a new layer of knowledge across domains. This is AI + X, great for building AI native, new applications. It is necessary, well-designed, and long overdue. But it is also only half the story.

To understand what is missing, one has to look closely at how other technological shifts have unfolded.

When nanotechnology emerged as a field, it evolved along two distinct trajectories. The first was a top-down approach, where existing structures were refined and miniaturised like transistors shrinking from visible components to nanometre-scale devices. The second was bottom-up, where entirely new materials were constructed from atomic building blocks, producing innovations that could not have been possible by top down refinements.

AI + X and X + AI mirror this distinction. They are not the same process viewed from opposite directions, but fundamentally different challenges that require different institutional responses.

The asymmetry that matters

In M, addition is commutative, meaning that X + Y is equal to Y + X. In artificial intelligence, this symmetry does not hold, and the difference is consequential.

The harder problem is X + AI. It is not about introducing AI to new learners, but about enabling those with years, often decades, of hard-earned expertise to integrate AI into how they think and act. In essence, AI + X only changes learners’ state, while X + AI transforms the state of knowledge itself.

Consider, for instance, the engineer who has spent a decade understanding the behaviour of materials at microscopic scales and knows that AI could compress months of simulation into hours, yet finds no structured pathway to bring that capability into her work. Or the water scientist whose long engagement with a specific river basin remains disconnected from computational tools that could make that knowledge predictive. Or the clinician with decades of diagnostic experience, such as a radiologist identifying early-stage cancers, who should be shaping how AI systems are used rather than adapting to them from the outside.

In each of these cases, the domain expertise is deep and irreplaceable. The tools are available, however without the right expertise of AI the outputs risk being misinterpreted or applied incorrectly, especially in specialised domains where errors are not always immediately visible. What is missing is the bridge between domain expertise and AI capability in practice and the ability to integrate both in real-world decision-making.

Across India’s research institutions, hospitals, and engineering systems, it is this missing bridge that may well be the single largest constraint on meaningful AI adoption. Current policy, for all its strengths, does not yet address this.

The numbers make this gap visible. According to the Global Skills Report by Coursera, India today has approximately 1.3 million AI learners, the highest anywhere in the world. Yet, it ranks 89th out of 109 countries in AI proficiency. More people are learning about AI here than anywhere else, but far fewer are able to apply it meaningfully in real-world contexts.

Why skilling cannot solve this but universities can

It is tempting to interpret this gap as a skilling deficit, but that diagnosis is not correct.

Skilling initiatives teach tools, enabling professionals to use AI platforms, run models, and interpret outputs efficiently at scale. These are valuable capabilities. But X + AI begins earlier, at the level of problem-framing and deciding what questions are worth asking, which data matters, and how to recognise when an AI-generated answer is subtly but critically wrong. These judgments depend on deep domain expertise and cannot be acquired through modular training alone.

They emerge instead from sustained engagement at the intersection of domain knowledge and technology understanding.

This is precisely where technology universities have a unique role to play.

At their best, these institutions bring together researchers who understand both the science of AI and the complexities of the domains it must enter. They are capable of engaging with domain experts not as trainees, but as intellectual peers working on shared problems. A senior scientist with decades of experience is unlikely to subordinate her expertise to a platform course, but she will engage seriously with an institution that has credibility in her field.

The government, interestingly, has already recognised this logic in part. The IndiaAI Mission’s Centres of Excellence in healthcare, agriculture, sustainable cities, and education are located within academic institutions precisely because meaningful domain-AI integration requires research depth. What has not yet followed is the extension of this logic across the broader system.

What must change

What X + AI demands is a different kind of institutional commitment, one oriented not toward first-time learners, but toward professionals deepening their practice.

This means designing programmes that begin with the realities of specific domains like energy systems, water management, healthcare delivery, agriculture, manufacturing and building AI capability into how decisions are actually made within those fields. It means moving beyond generic AI instruction toward engagement with real problems, real data, and real constraints.

As a first step, this requires universities to stop treating AI as a standalone add-on and instead embed it into the core of how disciplines are taught and practiced. This begins with their own faculty. Institutions must enable non-AI researchers and teachers to integrate AI into their domain workflows.

One way to do this is by creating domain-specific AI studios, where faculty, researchers, policymakers, and practitioners work together on real-world problems, combining domain knowledge with AI capability. Such spaces allow integration to emerge from practice rather than instruction. Creating such opportunities need resources and must be funded generously.

This shift must extend into the classroom. Instead of isolating AI within technical electives, it should be embedded into existing subjects like materials science students predicting properties before experiments, civil engineering students modelling traffic systems, medical students evaluating AI-assisted diagnoses, or policy students analysing legislation with AI tools. Assessment must move beyond technical use to emphasise problem framing, critical interpretation, and integration with disciplinary expertise.

Finally, universities will need to realign incentives. Cross-disciplinary teaching, dataset creation, and real-world deployment must be recognised and rewarded. X + AI becomes meaningful only when institutions are structured to improve how knowledge is applied in practice, not simply to expand AI literacy.

India’s AI education strategy is laying a strong foundation, and it deserves recognition for both its ambition and its execution. But a foundation, however well constructed, does not by itself create connection.

The bridge between India’s deep reserves of domain expertise and the transformative potential of AI will not be built through skilling platforms or curricular additions alone. It will require institutions that can hold both sides of that equation at once.

Technology universities are uniquely positioned to do so if they choose to take on that responsibility.

(The author is the Founding Vice Chancellor of Plaksha Universityand former Deputy Director of the Indian Institute of Science, Bengaluru)