Complex environments
"It uses deep learning and foundation models trained on large datasets of human driving behaviour [...] This means that vehicles are designed to benefit from vast amounts of real and synthetic driving behaviour data to safely navigate complex environments..."
I bet they haven't trained it on any of the roads around here. In recent times I have had to deal with...
- A car reversing back around a roundabout towards me, presumably having missed its exit.
- A peacock crossing the road in front of me.
- A bollard in the centre of the road where I had to drive round the wrong side of it because someone had parked so badly that I couldn't fit through on my side.
- A longish segment of road where parked cars leave only one car width, so if there's a vehicle wanting to come the other way, you and the other driver have to do a to-and-fro negotiation as to who goes and who waits.
And those are just a few examples - something unique and unexpected seems to happen fairly often, and if the training data doesn't cover it, who knows what the AI would decide to do?