A company tests a new self-driving car.
The AI has been trained with millions of road images.
During tests the vehicle drives very reliably.
It easily recognizes traffic lights, signs, and pedestrians.
But one day something strange happens:
On an empty road the car suddenly brakes hard and stops.
The passengers cannot see any obstacle.
After investigation it turns out
that a harmless object at the roadside caused the reaction.
Question:
Why can an AI in a self-driving car sometimes react to completely harmless objects?
Solution will be unlocked tomorrow.