This seems reasonable to me: if you do things that statistically correlate to being more likely to cause an accident, why shouldn't the insurance company charge you more? And since it's a relatively small group of drivers who cause the most accidents, if we could identify them the majority would see their rates drop. It's an interesting topic, and it would be nice to see more details on the accuracy of the algorithm,…
The question is, though, does "hard braking" actually correlate to being more likely to cause an accident? I don't think anyone has supported these assertions with evidence.