Insurance companies like to slice their customers up into buckets of similar risk and charge more or less accordingly.

Insuring an automated car will be cheaper. Insuring a manually operated car will be more expensive.

Sure, it will be more expensive than the automated car. But why would it be more expensive than today? Why would automated cars on the road make manually operated cars more risky than they are today?

At the moment, there is no other option other than human driving cars. Once automated cars are a realistic option, insurance companies can properly price insurance for human drivers much higher (comparatively speaking). A self-driving car should have almost 0 risk, meaning the vast majority of cars (if self driving) would carry no risk to the insurer.

> Once automated cars are a realistic option, insurance companies can properly price insurance for human drivers much higher (comparatively speaking).

What's "proper" about raising the prices for no reason? If they start pricing human drivers higher than they currently do, some other insurer will just undercut them.

yes, but that doesn't explain why the underlying risk of insuring human drivers would rise. Nor does it explain why, since the underlying risk is the same, Progressive wouldn't offer lower rates than Geico's.

I suspect that the risk would actually lower for the remaining human drivers as the AI vehicles are probably going to be better at predicting and dodging our stupid mistakes, resulting in fewer accidents even if we don't improve our driving abilities.

I don't think that those that choose to drive would provide more risk than they do today, however, they would provide (my gut and no more) substantially more risk than driverless cars. What incentive would an insurance provider have to insure a human driver? Much higher premiums I suggest.

The same incentive they have today: profit based on taking in more in then they have to pay out based on risk. If insurance companies arbitrarily bump up insurance for manually driven cars, then that opens up a market for someone to jump in and accurately charge based on risk.

You presuppose that American insurance companies are not a very effective cartel, primarily functioning in the regulatory capture industry, with a side business in insurance.

It would be cool if what you describe actually happened, but in the real world, premiums only ever move long-term in one direction: up.

More or less, I agree. With fewer accidents and fewer drivers insurance companies will be looking at decreased revenues. No company wants decreased revenues of that sort. Now, I'm not 100% familiar with the insurance industry and any regulatory capture that may take place, but then would be the time for a cartel to flex its muscles and ensure that startup insurance companies don't undercut them with fair pricing.