Even 10% less grievous bodily harm would be a massive improvement.

The fastest way to do that is to put size limits on personal vehicles. There has been study after study showing that the larger and heavier cars get the greater the risk to pedestrians. Pedestrian fatalities have almost doubled over a decade.

https://arstechnica.com/cars/2023/08/how-cities-can-stem-the...

> Pedestrian fatalities have almost doubled over a decade.

Requiring cellphones to disable themselves in cars might be even more effective.

What about the studies on deaths and injury with smaller vehicles only involving another vehicle? My guess is that it's much higher. Most families with kids want larger vehicles for exactly this reason.

Outperform mediocre humans means AI cars replace something shitty with something less shitty.

That means the net danger of traffic as a whole goes down.

Setting an unreasonably high bar - like "drive with perfect record" - will simply prevent that happening, and we're stuck with humans. Perfection is a north star, not an attainable goal. As long as we get net positive, the world is a tiny bit better. And then you make it a tiny bit better again. And again.

Why do you think the bar for AI should be higher than "better than what we have"?

The problem is that the bar for autonomous cars is perfection as far as public perception goes.

Right now, any incident involving a autonomous car is going to immediately be blamed on the autonomous car and not any other party, even if the fault lies entirely on the dickhead running the red light.

It won't be enough for autonomous cars to merely not cause crashes, people will expect them to actively avoid them, even in cases where a human would have failed.