The level 5 of autonomy is defined by the fact that the car can handle those new and extreme situations. I think that what you mention would match level 4 autonomy.

For example, definitions from https://www.techrepublic.com/article/autonomous-driving-leve...:

> Level 4: This is what is meant by "fully autonomous." Level 4 vehicles are "designed to perform all safety-critical driving functions and monitor roadway conditions for an entire trip." However, it's important to note that this is limited to the "operational design domain (ODD)" of the vehicle--meaning it does not cover every driving scenario.

> Level 5: This refers to a fully-autonomous system that expects the vehicle's performance to equal that of a human driver, in every driving scenario--including extreme environments like dirt roads that are unlikely to be navigated by driverless vehicles in the near future.

In my experience with machine learning, the last few percent of weird cases are exponentially varied and change over time so are always new. You hit a performance wall at a certain level. Researchers have been driven mad by this wall. See for example https://www.wired.com/2008/01/ff-aimystery/

For cars, behaving dangerously just 0.1% of the time is way too much.

A good AI should have really good 'one shot' deduction capability not just work by induction using tons of data. I don't think the current crop of algorithms are there yet.

Problem is that it is not very good pr if somebody dies or hurts badly every time cars encounter a "new" situation.

“Very rare” doesn’t cut it for modern safety standards.