Sure there is. Data on accidents involving human drivers is statistically useful for estimating risks from self-driving cars. Yes, driving machines are likely t

Sure there is. Data on accidents involving human drivers is statistically useful for estimating risks from self-driving cars. Yes, driving machines are likely to be better at driving than humans. But there a ton of other factors which are the same. The cars are the same shape, size, weight and materials as cars driven by humans. They drive on the same roads, in the same weather with the same passengers. It's true that we don't yet know how well self-driving cars will perform at scale. To say that we have no idea what risks are associated with them, or that they pose no risk to anyone but the passengers is ridiculous.