I think those statistics are difficult to get a complete picture on.
In california at least the manufacturers have to log miles driven in autonomous mode and the number of disengagement events over those miles.
That can be misleading though, since a car could log thousands of miles on a nice smooth stretch of interstate over and over, then spend significant amount of time disengaged in areas with pedestrians, cross walks, and stop signs.
The same pinch of salt should be applied to stats that someone like Uber themselves makes claims on, about how many miles they have driven autonomously with no incidents.
We could probably draw conclusions about safety per mile travelled on highways but probably not in trickier surface street scenarios without a lot more detail from the manufacturers.
In california at least the manufacturers have to log miles driven in autonomous mode and the number of disengagement events over those miles.
That can be misleading though, since a car could log thousands of miles on a nice smooth stretch of interstate over and over, then spend significant amount of time disengaged in areas with pedestrians, cross walks, and stop signs.
The same pinch of salt should be applied to stats that someone like Uber themselves makes claims on, about how many miles they have driven autonomously with no incidents.
We could probably draw conclusions about safety per mile travelled on highways but probably not in trickier surface street scenarios without a lot more detail from the manufacturers.