Beyond a campus ChatGPT subscription: what colleges actually need

Most Indian colleges meet AI the same way: a few faculty members pay for a consumer subscription themselves, find it genuinely useful, and the question reaches management. Should the institution buy this for everyone?

Per-seat pricing is the first thing that breaks. A mainstream assistant sits somewhere between ₹1,800 and ₹5,000 per user per month at list price. For a college of 2,500 students and 180 staff, licensing everyone is not a budget line — it is a capital project. So institutions license a handful of administrators instead, and the tool never reaches the people whose work it would actually change.

The second thing that breaks is grounding. A general assistant does not know your regulation, your syllabus, your examination pattern or last year's circular. It produces confident, plausible, unattributable text. For a student asking which unit an internal covers, plausible is worse than useless — it is wrong in a way that is hard to detect.

The third is disclosure. The moment an assistant becomes useful for institutional work, institutional material starts flowing into it: examination policies, internal notes, student records, accreditation drafts. Under the Digital Personal Data Protection Act, the college is the data fiduciary for that student data. Consumer terms of service are not a consent framework, and a policy page is not an architecture.

What institutions actually need is unglamorous by comparison. Answers grounded in their own documents, with the source citable. Access scoped by role, so a department's confidential material stays in that department. An audit trail. A price that does not grow with headcount. And an arrangement where the institution, not the vendor, ends up holding the index.

None of that is an argument against consumer AI. It is very good at what it does. It is an argument that an institution buying for 2,680 people is solving a different problem than an individual buying for themselves, and should stop shopping in the same aisle.