There are so many issues here.
First, we get to hear about the incidents where auto-driving (AD) leads to collisions, but very rarely where auto-driving prevents those collisions.
Second, human drivers lead to so many collisions that they are rarely reported. The evidence shows that AD systems are safer than human drivers.
Those two facts mean that AD-related collisions are over-reported.
The key issue is that human drivers routinely and regularly break the rules of the road, but in minor and socially-acceptable ways.
AD systems have road rules programmed into them and they obey those rules when human drivers would choose not to do so.
There are hundreds of reports of low-speed rear-end collisions in which an AD vehicle stops at a junction and the human driver behind rams into the rear. These happen because although the rules are that one should stop at junctions, many drivers routinely look to see if the way is clear as they approach and then do not stop.
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0184952
Something similar happened with the motorcyclist, according to the official accident report:
https://www.dmv.ca.gov/portal/wcm/connect/1877d019-d5f0-4c46-b472-78cfe289787d/GMCruise_120717.pdf?MOD=AJPERES
The AD vehicle was under AD control at the time.
ANyone who suggests it was not is contradicting the official accident report (above). Those reports are a legal requirement in California of any incident involving an AD vehicle.
The GM vehicle was preparing to move into a gap in the adjacent lane when the vehicle in front (in the adjacent lane) decelerated sharply, making the lane-change manouvre unsafe, so the AD braked sharply and moved back to its original lane. By this time a fast-moving motorcycle had moved into the area previously occupied by the AD car. As the AD car returned to its original lane, it side-swiped the motorcyclist.
I think this is a case where the motorcyclist has a legitimate case. The motorcyclist was travelling at around 17 mph - well within the speed limit.