Murder on the self-driving shared ride
App records and facial recognition proved who the three passengers in the car had been, where it had picked them up, and where they had intended to go. GPS logs and multiple traffic cameras (not to mention less traditional surveillance systems) showed where the car had gone instead, and how fast it had been moving when two of the passengers killed the other one by pushing her off the car.
The questions that gripped the Internet were others: Had they been justified in doing so, as the car had threatened to stop its out of (human) control speeding by crashing into a wall at full speed unless they killed the woman? And who had hacked the car to get her killed, and why?
The first question became exponentially more urgent a few hours later, when it turned out that it had been a worm, not a hack, and hundreds of passengers who hadn't been following the story --- or had preferred the new risk to the traumatic class demotion of public transit --- found themselves in the same position.
Proving statistical assumptions correct, few of them made a different choice.