two completely different systems (human and algorithm) working together to prevent crashes
Good point. Most current driver assistance systems perform terribly without human supervision.
Two different safety systems working together is always going to be better than one.
That might be true now, but I’m not sure it will always be true. Current autonomous vehicles aren’t quite as good as human drivers yet, so “defer to the human” is the safer choice. Once AV capabilities surpass humans, “defer to the human” might not be the safest choice any more.
I mean it’s always going to be better if they’re not working against each other. For example, imagine you’re learning to walk the tightrope. A harness and a net is going to be safer than either one. The harness could fail, and the net could have a structural weakness, but there’s very little chance of both happening. Or for a more car relevant example, crumple zone plus airbags is gonna help more with head on collisions than just one. When the two systems are in conflict, though, you’re right.
Good point. Most current driver assistance systems perform terribly without human supervision.
That might be true now, but I’m not sure it will always be true. Current autonomous vehicles aren’t quite as good as human drivers yet, so “defer to the human” is the safer choice. Once AV capabilities surpass humans, “defer to the human” might not be the safest choice any more.
I mean it’s always going to be better if they’re not working against each other. For example, imagine you’re learning to walk the tightrope. A harness and a net is going to be safer than either one. The harness could fail, and the net could have a structural weakness, but there’s very little chance of both happening. Or for a more car relevant example, crumple zone plus airbags is gonna help more with head on collisions than just one. When the two systems are in conflict, though, you’re right.