The truck is already on the ground and the vast majority of failures leave it there.
Does it matter? Many, many people are also on the ground in the proximity of streets, so if it goes wild it is still likely to hit pedestrians and vehicles. Drones could feasibly avoid busy streets by taking routes over rooftops for example; trucks simply cannot thereby forcing failure modes where people, cars and trucks collide.
Trucks are also much more expensive and contain even more valuable human drivers, this creates a much bigger incentive to keep things working safely. There's also more opportunity for the human driver to mitigate mechanical failures.
Agreed on the value disparity, but it seems to be in line with the opportunity for the vehicle to cause damage. Absolute worse case is a drone hits a person on the head and kills them, but statistically those odds have to be remote compared to other non-fatal injuries or property damage -- certainly when compared to the opportunity for injury and property destruction caused by a truck.
The human driver also provides a significant failure mode (medical problem, distraction, driver error), which I suspect is a much larger risk than the small chance the driver could use to avert accidents after a mechanical failure.
As for maintenance a lot of businesses operate very close to the margin, sometimes in the red. They're going to save money everywhere they can, this includes doing the absolute minimum maintenance and running every drone until complete failure, it just becomes a question of how they fail.
True, but doesn't that also suggest that is the current operating mode for truck fleets? Isn't it overall better to have vehicles with less kinetic energy, even when poorly maintained, from a safety point of view?