While hydrogen-air mixes explode really readily (outstandingly wide flammability and pretty wide detonation range), and the energy released is considerable for the weight, the actual explosion does not produce a particularly high overpressure wave.
That's because the starting density of the hydrogen air mixes at near atmospheric pressure (such as in a balloon) is pretty low. Also, the balloon does not significantly contain the explosion, which reduces the danger a lot. I would not want to do it in a glass container.
My chemistry teacher told us how once when he ignited helium in a test tube, the tube broke and he ended up with pieces of glass embedded in his skin. The students had face masks and he was looking the other way "just in case" for this "safe" experiment but he could have easily been blinded.
Things can always go wrong. We probably shouldn't strive for 100% safety because they we'd spend our lives in a padded cell. But we also shouldn't assume things are safe because they're common or routine.
Along with the other commenter, I'll add that a classroom is usually a lot bigger than a home dining room or other domestic party locations. That size also helps things dissipate instead of reflect. Not sure by how much but I'm sure it does something.
