Thoughts: LLMs Are a Multi-Billion Dollar Industry, but Unis Aren’t Ready for Them
TL;DR: An international student dumping some thoughts on universities and AI use, and curious about what others think.
Edit: I've compacted my arguments so that it more reflects what I really wanted to express
So last year in a Biology lecture, I heard the lecturer said: “ChatGPT gives wrong answers because it hallucinates.” He then showed a pretty specific biology question pasted into Chatgpt to prove the point. Today, in another biology workshop, I heard almost the exact same argument again from a tutor. However I feel like this understanding of AI is still stuck in 2024.
LLMs are fundamentally probabilistic prediction machines. “Hallucination” is real, but so is “garbage in, garbage out.” If you give bad prompts, no context, and don’t verify anything, obviously the output will be unreliable. The problem is that universities often respond with blanket restrictions instead of teaching students how to use these systems properly. So students either: 1) panic and pay random “AI humanizer” websites, 2) or become ashamed of using AI at all, while their understanding of AI never moves beyond “getting ChatGPT to write an essay.” But AI isn't just about chatbots. There are entire systems around it now: agentic workflows, tool use, context engineering etc. research pipelines are also being seriously discussed in academia. This creates a strange gap between the ivory tower and the outside world.
Another thing I find surprising is that there are still almost no undergrad subjects seriously engaging with AI itself outside traditional CS pathways. Meanwhile Stanford already has subjects like CS146S. For a conservative university, maybe you could still say “it’s too early.” But if universities still have no meaningful engagement with these topics by 2027 or 2028, I honestly think that's a problem.
As an international student, this feels frustrating. Sometimes it feels like we're paying enormous tuition fees while studying inside a system that is slowly becoming disconnected from the industries. I especially feel this in some education/humanities breadth subjects where grading criteria can become extremely subjective and difficult to justify transparently. I’m not a software engineering student, so maybe some of my concerns are incomplete or wrong. But I do think that if students don’t actively explore these technologies themselves, universities are unlikely to reform proactively. And honestly I suspect the next 5 years are going to change higher education much more dramatically than many institutions expect.
Just some of my random thoughts. And I'm curious what others might think.