If the SF Chronicle actually wanted to inform their readers they would publish a table of all motor vehicle crashes in the city by type of vehicle.
This isn't a simple matter of "humans good drivers, robots bad drivers"; it's an example of "humans can improvise and respond reasonably to novel situations; robots cannot reliably interpret and respond to novel situations."
You claim “ There are few vehicle-pedestrian crashes where the offending vehicle parked on top of the vehicle and wouldn't be moved” - what is your evidence for that? 2.3 million people are injured in car accidents in the United States per year.
Finally, it is plausible that AVs will have a higher frequency of certain very specific failure modes relative to human drivers, while having dramatically lower overall frequency of accidents where people are injured. This type of reporting obscures that.
> While spokespeople for the Police Department did not have an officer’s report Tuesday morning to clarify exactly when police made the order, they said it is standard procedure not to move cars at the scene of wrecks — regardless of whether a human is operating them.
They highlight the helplessness and cold hard reality of being at the wrong end of a miscalculation.
So even if the overall number of accidents is less, the number of times a human driver would park on top of a victim of a hit and run and turn on the hazard lights is near zero.
At some point it won't be news when driverless cars are involved in accidents but for now they are. Personally, I think the public should be informed about what a new technology is doing and capable of whether it's good or bad. Especially one that can kill people rather easily.