> Insurance companies may think you’re a higher-risk driver if you’re accelerating fast, braking aggressively, making sharp turns, using your phone while you drive, coming home at 3 a.m., taking dangerous roads or simply driving a lot.

Is there any proof that any of these behaviors actually correlate with tangible risk? (I mean, I can understand that "simply driving a lot" increases risk.)

Personally, I'd be more interested in cars with cameras using machine learning to build a risk profile of other cars. I'm not the only person who sees dangerous behavior all the time. It would be interesting if I could sell a stream from my cameras to a company that builds risk profiles based on observed behavior.