Love your content, Phil. I agree that the minimum insurance requirement should be higher—not from a premium affordability standpoint, but to ensure financial stability across the system. In a serious accident, the courts will determine liability. If the damages exceed the insurance coverage, the liable party—typically the AV fleet operator—must cover the shortfall. While this isn’t a major issue for companies like Waymo or Tesla (both of which own insurance captives), many AV startups run on thin balance sheets. A nuclear verdict involving multiple fatalities could easily bankrupt them, leaving victims without adequate compensation. Bottom line: a higher insurance minimum protects both vulnerable startups and accident victims, though it's less critical for well-capitalized players with fortress balance sheets.
As someone who has been in an accident where there was an injury ( even when not at fault) I can tell you that the current minimums are way too low. I’m glad I had insurance above and beyond the minimum.
This topic can be misleading. There is no federal mandate on the value of VSL to be used in the private sector during Hazop, SIL or similar risk analysis. The government doesn't use VSL to determine what is "R" in ALARP.
More importantly, it is not used to determine maximum insurance for human drivers
So there is no real reason why it would have to be mandated specifically for autonomous drivers.
I agree there is no current mandate to base insurance, court judgments, or HARA on VSL.
However, VSL is a justified number that some would consider relevant to those use cases. Removing human drivers from being responsible for driving safety is a huge disruption to the status quo. The question is, what are the pro/con for adopting VSL as part of that change?
Also, in the US it is not ALARP, but rather cost effectiveness. If there is systematic under-valuation of human life (which is prevalent), then the bar is set low for safety improvements.
Saying it is not how things work right now is one argument, but not the only possible argument.
I support correcting the under-valuation of human life.
But I don't think it's fair to single out autonomous driving companies and force on them a safety standard much much more stringent than what is enforced for oil&gas, process industries and other high risk scenarios (including human driving).
I believe the DOT VSL number is close enough to other regulatory settings as well (EPA, FEMA). Enforcement in other industries would be a good thing too. (Important not to succumb to whataboutism here.)
The personal vehicle problem is a bit different. The US car industry has waged a 100-year-long campaign of car dependence, and successfully lobbied against being as safe as driving is in many other advanced companies. And pressured for low liability amounts (e.g., the "ambulance chaser" and "nuclear verdict" tropes). US Individual drivers have to pay for insurance, and so are caught in a tough spot that we should work on (unaffordable premiums due to higher crash rates, owning a car required to survive, no viable public transit many places), but that is a systemic problem that will take decades to unwind when we want to start (which we don't yet).
But that does not mean that companies chasing a trillion dollar market by promising their technology is vastly safer than people should be given a discount on the harm they do. Let them put their money where their mouths are.
More generally, I'm observing the VSL to insurance minimum disconnect and think it interesting to see people's points of view, so thanks for the discussion.
Love your content, Phil. I agree that the minimum insurance requirement should be higher—not from a premium affordability standpoint, but to ensure financial stability across the system. In a serious accident, the courts will determine liability. If the damages exceed the insurance coverage, the liable party—typically the AV fleet operator—must cover the shortfall. While this isn’t a major issue for companies like Waymo or Tesla (both of which own insurance captives), many AV startups run on thin balance sheets. A nuclear verdict involving multiple fatalities could easily bankrupt them, leaving victims without adequate compensation. Bottom line: a higher insurance minimum protects both vulnerable startups and accident victims, though it's less critical for well-capitalized players with fortress balance sheets.
Yes, this is an excellent point.
I think one solution is to make functional safety a lot cheaper, avoids recalls and drives down costs.
As someone who has been in an accident where there was an injury ( even when not at fault) I can tell you that the current minimums are way too low. I’m glad I had insurance above and beyond the minimum.
This topic can be misleading. There is no federal mandate on the value of VSL to be used in the private sector during Hazop, SIL or similar risk analysis. The government doesn't use VSL to determine what is "R" in ALARP.
More importantly, it is not used to determine maximum insurance for human drivers
So there is no real reason why it would have to be mandated specifically for autonomous drivers.
I agree there is no current mandate to base insurance, court judgments, or HARA on VSL.
However, VSL is a justified number that some would consider relevant to those use cases. Removing human drivers from being responsible for driving safety is a huge disruption to the status quo. The question is, what are the pro/con for adopting VSL as part of that change?
Also, in the US it is not ALARP, but rather cost effectiveness. If there is systematic under-valuation of human life (which is prevalent), then the bar is set low for safety improvements.
Saying it is not how things work right now is one argument, but not the only possible argument.
Valid points.
I support correcting the under-valuation of human life.
But I don't think it's fair to single out autonomous driving companies and force on them a safety standard much much more stringent than what is enforced for oil&gas, process industries and other high risk scenarios (including human driving).
I believe the DOT VSL number is close enough to other regulatory settings as well (EPA, FEMA). Enforcement in other industries would be a good thing too. (Important not to succumb to whataboutism here.)
The personal vehicle problem is a bit different. The US car industry has waged a 100-year-long campaign of car dependence, and successfully lobbied against being as safe as driving is in many other advanced companies. And pressured for low liability amounts (e.g., the "ambulance chaser" and "nuclear verdict" tropes). US Individual drivers have to pay for insurance, and so are caught in a tough spot that we should work on (unaffordable premiums due to higher crash rates, owning a car required to survive, no viable public transit many places), but that is a systemic problem that will take decades to unwind when we want to start (which we don't yet).
But that does not mean that companies chasing a trillion dollar market by promising their technology is vastly safer than people should be given a discount on the harm they do. Let them put their money where their mouths are.
More generally, I'm observing the VSL to insurance minimum disconnect and think it interesting to see people's points of view, so thanks for the discussion.
https://en.wikipedia.org/wiki/Value_of_life