Phillip Gingras

How do we reframe AI thinking in Higher Education?

Something that I’ve experienced for the last year has been the rise of everyone’s readiness to be “AI first” whether that’s tool to automate syllabi or to review academic integrity. More and more people want their AI tools to do things, rather than create text or cite sources. But, one fundamental problem is, we want agentic action in an LLM world.

It’s reminiscent of the Blockchain boom years back where everyone sought to find ways to implement blockchain technology as soon as was humanly possible only to realize that while it has great applications, those applications are rather circumscribed. It’s the same with AI. There’s a full-tilt sprint to adopt AI, but everyone is buying LLM tools, most of which are wrapper around tools buit from models developed by one of the three hyperscalers. LLM is great at what it does, but it’s incredibly circumscribed.

AI is certainly the sum of its parts, and that includes ML, GNN and Agentic. Higher Education is unique in that it’s a data rich environment that could benefit from machine learning and agentic AI. From student success systems to complaince and accreditation, there are use-cases, but everyone is so focused on the LLM and classroom that I feel we’re missing the broader, more important, applications of AI.

So far, it’s just a thought.