I was recently asked by an HEI what I thought about a recent ‘AI enhancement to a student records system’. It appears as though every technology vendor pitch landing on a university leader’s desk this year claims to be “AI-powered.” The phrase has become almost meaningless.
A chatbot drafting enrolment emails and a system flagging irregular exam behaviour might both carry that label. They share almost nothing else: different training, different oversight, different consequences when they get something wrong.
So, the experience has prompted me to publish a background briefing setting out, in plain language, the six broad categories AI now takes, from generative chatbots through to the older, more transparent rule-based systems still quietly running your compliance checks. The point isn’t to make senior leaders technically fluent. It’s to give them the vocabulary a procurement conversation actually requires, along with a short glossary and a set of questions worth putting to any vendor before a contract is signed.
The distinction that matters most: capability and explainability move in opposite directions. The more open-ended and powerful a system is, the less anyone, including the vendor, can say with confidence why it produced a particular answer. Regulators have already worked this out. Higher education’s own procurement processes mostly haven’t.
One line from the research has stayed with me: “beware of bright shiny objects.” Good advice, and hard to imagine a vendor volunteering it unprompted.
None of this is an argument against buying AI. It’s an argument against buying it without first making the vendor say, plainly, which category it actually is.
Before your next vendor meeting: does your team currently ask that question, or does “AI-powered” still get a pass?
