Could self-driving cars made by different companies crash into each other?
I’m writing a romance novel that is slowly becoming more and more sci-fi.
The book takes place a little in the future when there are multiple models of autonomous vehicles (AVs) on the roads, each with their own strengths, weaknesses, and ‘personalities.’ At the beginning of the book, we learn that there has been an uptick in crashes between competing AV models.
My current explanation for this is that the cars would have been trained on the behavior of human drivers—not other AV models—leading them to be less skilled at predicting the behavior of those other models (especially those that are newer).
But is that even realistic??? I know that when vehicle-to-vehicle communication software works properly, cars can communicate nearly instantaneously. But is it reasonable to believe that V2V communication might not be able to fully prevent these crashes?
(maybe due to the spontaneous nature of certain road conditions, lags in signal transmission, the size of the data that would be necessary to actually prevent the crash, etc.)
Ultimately, I want the messaging of this novel to be about the importance of transparency in AI. I’m just worried that the solution to the problem I’ve outlined is more technical than it is ethical. Any help is much appreciated!