Nonsense. Uber's self-driving cars currently have something like 30x the fatality rate of an average driver and that is in cherry-picked locations that they chose for testing and with a human driver constantly watching it to take over when it does something wrong. There is no evidence that they are anywhere close to surpassing human drivers.
Given a sample size of 1 I don't think there's much to be discerned form the data at all either way.
> There is no evidence that they are anywhere close to surpassing human...