Maybe you can't prevent an accident, but you can at least reduce momentum and demonstrate that your system was working correctly.
In my driving lessons I got told to swerve for humans. Pedestrians and cyclists are the least protected participants in traffic. If you swerve and hit another car, the resulting crash will cause a lot of damage, but injuries to humans will likely be much less severe, if they even happen.
Animals, especially small ones are a different matter. Using the same argument, it is better to risk hitting the animal than risking sverving into other traffic.
The question isn't about the MHZ speed of a computer or the theoretical reaction time but the lack of reaction by the computer in the actual real world. It doesn't matter if this is a sensor failure, a software bug, or a slow reaction by a computer. The result is the same. A person died.
From the article:
*"... the lack of braking or swerving whatsoever is alarming and suggests that the system never anticipated the collision.”
https://blog.prototypr.io/automation-is-making-us-dumber-and...
Of course, this all presumes you can quickly analyze the situation and conclude that it is safe to swerve (which, in theory, the computer can do far faster than a human can).