Why AI Belongs Inside a Sports Degree, Not Beside It

Most Indian institutions are adding AI to sports education as a certificate alongside the degree. That is the wrong shape. A note on what changed, what a curriculum actually has to carry, and the constraints — including one in law — that nobody plans for.

There is a version of this argument that I am not going to make. It is the one where sport is described as being transformed by artificial intelligence, a number is quoted about the size of the market, and the conclusion is that institutions must act now or be left behind. That argument is everywhere, it is mostly unfalsifiable, and it has produced a great deal of expensive, badly designed provision.

The narrower claim I want to make is this: the work a sports graduate does has changed in a way that makes data literacy a core competence rather than a specialism — and that a curriculum which treats it as a specialism will produce graduates who cannot do the job. The distinction matters, because it determines where the material goes.

What actually changed

Three things, none of them about AI as such.

The first is measurement. Positional tracking, wearable load monitoring and video analysis are no longer confined to elite setups. A state academy can now generate more data in a week than a national side had access to twenty years ago. Someone has to interpret it.

The second is that the interpretation is increasingly done by a model rather than by a person. Injury-risk flags, workload recommendations, opposition reports, talent-identification shortlists — these arrive as outputs. The practitioner’s job is no longer to compute them; it is to judge whether to act on them.

The third is that this judgement is now the daily work of ordinary roles, not specialist ones. A strength and conditioning coach at a district academy reads a load dashboard. A PE teacher is handed a fitness-assessment platform. A team operations manager is sent a scouting model’s output. None of them are analysts. All of them are now expected to evaluate something a machine produced.

The skill that matters is not building the model. It is knowing when the model is wrong, and having the standing to say so.

Why the certificate-alongside-the-degree model fails

The common response in India has been to bolt an analytics certificate onto an existing physical education or sports management degree. It is administratively easy: no curriculum revision, no new approvals, a visible line on the brochure. It also does not work, for three reasons.

• It separates the tool from the judgement. A student learns Python in one module and biomechanics in another, and is never required to use one to interrogate the other. The integration — which is the actual skill — is left as an exercise for the graduate.
• It makes the material optional in practice. An add-on is something a student can pass through. Core assessment is something they cannot avoid, and avoidance is exactly what you do not want with a competence this central.
• It signals that the department does not believe it. Faculty read the structure of a curriculum accurately. So do students. Material placed at the edge is treated as being at the edge.

If a capability is genuinely core to the work, it has to appear in the core of the curriculum — assessed, sequenced, and taught by people who also teach the sport. That is a harder institutional decision than adding a certificate, which is precisely why so few have made it.

What the policy framework already asks for

This is not a case that has to be argued from first principles in India. NEP 2020 is explicit about multidisciplinarity, about technology integration, and about physical education being treated with the same academic seriousness as any other discipline rather than as a co-curricular activity. The NCF-SE 2023 carries the same logic into school-stage curriculum design.

What the policy does not do is specify how. That is left to institutions, and it is where most of the difficulty sits. A framework can tell you that technology should be integrated. It cannot tell you whether your faculty can teach it, what it costs, or what you are allowed to do with a fourteen-year-old’s movement data.

The three constraints institutions underestimate

Having designed this into a four-year degree rather than beside one, these are the places it gets difficult. None of them are technical.

Faculty capacity is the binding constraint. There is no large pool of people in India who can teach both exercise physiology and applied data analysis. You are choosing between training sports faculty in data methods, training data faculty in sport, or team-teaching. Each is slow. Budgeting for software and ignoring this is the most common failure, and it is fatal: the curriculum exists on paper and is not delivered.

Cost is manageable if you resist the equipment pitch. A great deal of useful work runs on free tooling and open data — public match datasets, browser-based notebooks, open-source video analysis. Expensive tracking hardware teaches students to operate one vendor’s system, which is a narrower and more perishable skill than being able to reason about data from any source.

The children’s-data question is the one nobody plans for. If your programme places students in schools, runs a feeder academy, or assesses minors in any way, you are processing children’s personal data — and under India’s Digital Personal Data Protection Act, 2023, that is governed territory. Section 9 requires verifiable parental consent for processing the data of anyone under 18, and restricts tracking and behavioural monitoring directed at children.

There is an important nuance that most commentary misses: the Act and its Rules provide for exemptions, and educational institutions undertaking tracking or monitoring for educational purposes or child safety fall within the classes contemplated. So this is not a prohibition on assessing students. It is a requirement to know which basis you are operating on, to be able to show it, and to have designed consent and data handling before the first dataset is collected rather than after. Take this to counsel rather than to a vendor — and note that the operational detail sits in the Rules, which have moved, so check the current text rather than a summary.

I would put it more plainly: a sports department that teaches students to collect movement data from minors without teaching them the law that governs it has not finished designing the curriculum.

What a graduate should actually be able to do

Stripped of the vocabulary, three competences:

• Read a dashboard and say what it does not show. Most sports data is partial, and the gaps are where the mistakes live — a load metric that ignores sleep, a talent model trained on a population unlike the one in front of you.
• Interrogate a model’s output before acting on it. Not derive it — interrogate it. What was it trained on? What would make it wrong? What does it cost to be wrong in this direction rather than that one?
• Explain the answer to someone who does not want to hear it. A coach, a parent, a selection committee. An analysis that cannot survive that conversation does not change any decision, and an analysis that changes no decision was not worth producing.

None of these require a student to become a software engineer. All of them require that the data work and the sport science sit in the same room, taught together, assessed together.

What I am not claiming

Two honest limits, because a piece like this is worth less without them.

First, the evidence that AI-integrated curricula produce better practitioners is thin, because the approach is too new for outcome data to exist. The case I have made is a reasoned one about the shape of the work, not an empirical one about graduate outcomes. Anyone telling you otherwise is ahead of the evidence, and that includes anyone selling you a platform.

Second, there is a real risk of producing graduates fluent in tooling and weak in the underlying science — people who can generate a report they cannot evaluate. That failure mode is worse than the one it replaces, because it carries false confidence. Guarding against it means the physiology, biomechanics and pedagogy have to get harder alongside the data work, not easier to make room for it.

Where we stand

At Sportal, this is designed into the core of the B.P.E.S. and B.S.M. degrees awarded by Mind Power University, Bhimtal, rather than offered beside them. Our first cohort arrives in the 2027–28 session, so we have a curriculum position and an argument, not a set of outcomes. We will publish what we learn, including the parts that do not work.

I would rather be judged on that than on a brochure claim. If you are designing something similar at your own institution, or you think this argument is wrong, I am genuinely interested in the conversation — the research collaborations we run through the Erasmus+ SPIRIT project across India, Indonesia and Sri Lanka began as exactly that kind of disagreement.