A preliminary study from the University of Michigan Transportation Institute took a look at autonomous vehicle crashes reported by Google, Delphi, and Audi, all of which have licenses to operate self-driving vehicles in a number of states. The researchers compared that data to adjusted statistics pertaining to conventional vehicles. The study showed that self-driving vehicles were actually involved in more accidents on average, per million miles traveled, than their conventional forebears.
However, the researchers cautioned that those numbers may not tell the whole story. The data was pulled from 11 crashes among three makers of self-driving cars whose fleets only cumulatively drove 1.2 million miles on public roads. The data for conventional cars, however, was derived from a sample of the reported crashes that occurred over 3 trillion miles of annual driving. (The researchers did, however, adjust the data to account for unreported crashes.)
Because of the statistical uncertainty that comes with comparing a census of autonomous vehicle crashes with a sample of conventional vehicle crashes, the researchers couldn’t say for sure that self driving vehicles are more likely to be involved in crashes than conventional cars.