Driver age isn't the factor that explains the difference, accident rate adjusted for age shifts up a similar amount for both Autopilot and regular drivers. Wha

Driver age isn't the factor that explains the difference, accident rate adjusted for age shifts up a similar amount for both Autopilot and regular drivers. What truly makes a difference is the kind of road they're driving on. The accompanying paper provides a figure showing Autopilot is essentially only used on freeways which are inherently safer. Adjusting for that seems to explain most of the difference. Though personally I'd prefer to have just compared the accident rate on freeways for both types rather than this weird weighting method they try to use.