I am, in general, hoping AV will reduce road deaths in the future.
The last hurdle is regulatory. We can’t let AV manufacturers use “there’s no driver” as a way to escape responsibility, externalizing the harms AC cause onto society.
The question is how to achieve fairness. If a human driver commits vehicular manslaughter, they get the book. What about AV? $10 million? Executives go to jail? What if $10 million fine per X AV miles driven is an OK cost of doing business?
Who does it benefit if an accident ruins a second life?
What does a jail sentence deter? ("[no] gross negligence [...] wasn’t engaging in a race or sideshow, was not texting, and was not under influence")
This person was 80 years old with no criminal record, needs to pay $67400 in restitution, do 200 hours of community service, isn't allowed to drive for 3 years but "never intends to drive again". Apologised to the family of the victims. She's taking responsibility and I can't imagine forced labor at that age is fun. What more can you ask for here? The family member isn't coming back if she gets what's not unlikely to be a life sentence
Edit:
> She told a witness at the scene that she was trying to park her car when she accidentally moved her foot to the gas pedal.
This seems to happen a lot. Don't know about statistics but this happened to someone I know at 50yo (thankfully only damaged their own car minorly), and you hear it on the news with some regularity. Maybe the gas needs to be in a fundamentally different spot from the brake? We can jail the people to whom it happens, sure, but I can understand a judge using their head instead of their heart. The real solution must come either from the automotive industry or legislation
In the US, 11 deaths per billion miles driven (or about 47k per year) is currently seen as an OK cost.
More than twice as much per mile as places like Sweden and Switzerland, and still substantially more than places like Canada, Australia or Germany (all three in the 6-8 deaths per billion miles range). So it's not like there isn't room to improve. The effort to do so just isn't seen as worth the cost at the societal or government level
Turning that into a monetary cost would change the ethics slightly, but it wouldn't be a monumental shift
The issue here is that a lot of the concerns about AV's are orthogonal to the standard metrics of concern.
I'm a strong transit alternatives advocate, but even I recognize that a firetruck or ambulance being blocked by an AV has the potential to cause an outsized amount of death and destruction, because deaths aren't always linear and a fire that is able to get out of control can do catastrophic damage compared to a single out of control vehicle.
I'm genuinely stunned that AV's do not have the ability to be "commandeered" by Police/Fire/EMS in a pinch, and I'm honestly surprised that regular citizens can't just hit a red button that signal "this is seriously an emergency." These are fairly simple steps to mitigate the tail risk of AV's but the platforms aren't going to prioritize that if there are no incentives.
Don't forget to add rail incidents to that metric. I live in Spain, this year we had 4 derailments for a total of 48 deaths and 195 injured. The USA has had 0 passengers killed or injured from train accidents this year. Portugal had 15 death after a tram derailment. In Amsterdam, the tram is more dangerous than the car.
Also Germany is very high (for European standards) because of the Autobahn. They can save around 140 lives a year by having a limit on the Autobahn but the car lobby in Germany is very strong. Those 140 lives are seen as an OK cost just to go vroom on the Autobahn.
Coming from a bio background, I’ve always been confused why auto fatality stats are normalized per miles driven. Epidemiological metrics like incidence or prevalence seem like they would work fine? Town A would be “safer” than town B if people’s commutes are 20% shorter, even if accidents occur w same frequency
> So it's not like there isn't room to improve. The effort to do so just isn't seen as worth the cost at the societal or government level
That effort being what, exactly?
Road fatalities per mile driven don’t translate cleanly from country to country because the type of roads and even types of deaths (single vehicle, multi vehicle) are different.
We could set the speed limit at 25mph everywhere and force all vehicles to not exceed that limit and that would make the number go down, but the cost would be extreme for everyone.
So what, exactly, are the solutions you are proposing?
Doesn’t that 11 per billion statistic include commercial drivers as well? And doesn’t the United States have by far the largest percentage of commercial miles driven of any developed nation?
There’s a far cheaper solution available. Log books.
We subsidize driving by somewhat over a trillion dollars annually, mostly due to lax penalties for negligence which shift liability to drivers’ victims[1]. One way to tackle all of these problems would be requiring drivers to cover the full damages.
Another simple and effective measure would be changing fines from absolute values to a percentage of income. Right now, parking in a bike lane usually doesn’t kill anyone so drivers are only thinking there’s a small chance of a small fine, but if it was a chance of, say, 0.1% of annual income Waymo technology would magically be capable of not doing that. Add a right of private action and enforcement would be high enough to really speed things along, too, and that’d improve safety and travel times for all road users.
Yeah, making fines relative to income would change behaviors for sure. A $20 ticket when you make $20 an hour hits different when you're making $200 or $2,000/hr. If it was a percentage of pay, then the ticket would actually sting.
> We can’t let AV manufacturers use “there’s no driver” as a way to escape responsibility, externalizing the harms AC cause onto society.
There is essentially nothing to be gained from doing this because it will not in either case be manufacturer; it will be an insurance company.
If the liability is paid by the vehicle owner's insurance then things work as they do now. You buy a car, insure it, if there is a liability there is an insurance claim and then the victim has someone to pay them for their injuries. Meanwhile the manufacturers still have a financial incentive to make safer cars because buyers want neither accident prone vehicles as the one they use nor high insurance rates. The insurance rates in particular are in direct competition with the car payment for the customer's available income.
Whereas if you try to put the liability on the manufacturer, several stupider things happen.
First, they're just going to buy insurance anyway, but now the insurance cost has to be front-loaded into the purchase price, which increases costs because now you're paying car loan interest on money to cover insurance five and ten years from now, when you otherwise wouldn't have needed to pay the premiums until the time comes.
Second, what happens to cars from manufacturers who no longer exist? They can't continue paying for insurance if they're bankrupt, so you need it to be someone else. Worse, if a company produces a vehicle which is unsafe, that will tend to cause them to go bankrupt. But then people still have them, and would continue to operate them if they're allowed to point victims at the bankrupt manufacturer, whereas the incentive you want is for the premiums on those cars to go up for the vehicle owners so that they stop operating them.
> If a human driver commits vehicular manslaughter, they get the book. What about AV?
They get their licenses pulled statewide [1]. Cruise's single negligent manslaughter event carried more consequence than dozens of human cases combined.
I also hope AV will reduce road deaths in the future but I don't think what will make the difference is regulatory. Rather the tech will advance from doesn't work to works in Waymos but is expensive to works in most cars and has become cheap.
> What if $10 million fine per X AV miles driven is an OK cost of doing business?
It’s the same cost/benefit we accept under current rules. Why have cars that can go 3x the speed limit? Why not require breathalyzers in cars before starting them? Why not fine logistics companies if one of their drivers breaks the law? And so on… Because it’s worth it
Your questions are pertinent but what’s the benefit "worth" you’re referring to? The two first proposals would risk a politician popularity and the last one would be lobbied to he’ll buy the logistic companies. IMHO inconvenience isn’t worth driving among drunk coursier at 200kmh.
These people want to play god with our lives but at the same time move fast and break things. Look at software quality anywhere, it's a mess and only about to get much worse.
We should not let them. Jail time for anyone involved in any of the decision making process, applied at scale with the number of vehicles and deaths.
Why should the standards be any different? They want to change the status quo with tech only so they can get paid and extort us with yet more subscriptions.
AVs will never substantially reduce road deaths. They will optimize to just being slightly better than human, but fail in new and more unexpected ways. There is not enough incentive for them to make it safer.
If we consider fairness/retribution/justice then we won't get this future of less road deaths.
1. There will always be a probability of death from a vehicle. This can never go to 0%.
2. If the probability of a AV causing death is many magnitudes lower than human driving then that is the future we must choose.
If 1 and 2 holds and we hold AV manufactures accountable in the sense that Executives go to jail or are personal liable financially for deaths/injuries then AV will never get released or become mainstream even if this results in less total deaths. The sense of fairness/justice/retribution may make us feel better but result in more overall deaths. Logically this means that there must be a standard. Something like x deaths per y cars manufactured. If above the threshold you get big fines as a company. As technology gets better you can lower the threshold. Anything apart from causing deaths either purposefully or negligently would have be ignored.
Can we as a species accept this? That is another question.
We can look to other forms of automation to get a sense of what to do. For example, planes largely fly themselves and a loss of life due to manufacturing errors from the manufacturer would deem them liable for those deaths. Seems like the solution here is large penalties and generally broad disincentives for incurring harm.
Societies can already reduce road deaths to nearly zero, it's cheap, it's easy, and it's fun. It's just redirecting all of the cash we spend on vehicles/cars/highways/roadways/signs/etc into public infrastructure that is all encompassing.
A hundred billion dollars a year [0] on construction (reading the definition I'm not 100% sure what is included in this due to how definitions can be hazy) has goes a long way, not to mention the amount we spend on gasoline, car maintenance, etc etc.
The reason I say it's fun, is because I love being on a train. First time I was able to ride one, which due to living in the good old USA wasn't until I was 23, I yelled "I'm on a train" . The Germans traveling with me weren't as into it.
Just because you like trains does not mean that it is actually a solution to everyone’s problem. For example, until proper law-enforcement starts happening on public transit, nobody in my family is allowed to take it in USA (they are allowed to do so in Singapore or Japan)
The CEO gets charged with manslaughter? I work in healthtech and the responsible individual is certainly personally liable for any harm that results from reckless behavior, it should be the same here.
Same as if someone were driving, if a person just jumps in front of your car while you're driving under the limit/sober/etc, you aren't at fault, so the AV should also not be at fault if it couldn't reasonably avoid the harm. You balance these things, benefit to society vs harm to society, and you come to an acceptable tradeoff.
Could you provide examples of healthcare executives held personally liable for harm resulting from reckless decision-making? I have never heard of such a thing happening in healthcare so framing CEO responsibility as a solution to the problem sounds like a stretch to me.
Some examples: Elizabeth Holmes got canned for lying to investors, not harming patients. Purdue Pharma plead guilty to misleading regulators and giving doctors kickbacks, not causing some hundreds of thousands of opioid deaths, but no Sackler family members were personally tried.
> The CEO gets charged with manslaughter? I work in healthtech and the responsible individual is certainly personally liable for any harm that results from reckless behavior, it should be the same here.
This is in like China, yes? Certainly not in the US of A, hence Luigi and all that…
We've had public transportation for a couple of centuries and no where has it really led to a road death free utopia. I like public transport but no harm trying something new.
I had to look up a name for this. "Utopian Fallacy."
You don't have to get rid of genuine progress just because your utopian vision has something better. The USA is on the path to autonomous vehicles. They are not on the path to public transportation excellence.
Yup. Even if "safer per mile", more cars and more miles driven will probably outweigh the benefits. And still be hazardous to cyclists and pedestrians, still make us design stupid cities (built for cars, not people), etc.
Like how electric cars were for saving the car companies, not the planet, autonomous will be the same.
Simple. Blame the owner of the vehicle. They relied on automation and it failed. They go to jail for negligent homicide (whatever flavor is appropriate). That will tank sales of any AV tech that cannot maintain standards.
People are killed by industrial equipment fairly regularly.
I'd say we actually have a perfectly functional legal framework for all of this, and the real issue is a lot of new people are about to find out it also applies to them as well.
Whether it was working well in the first place is the real question.
Many things already reduce road deaths and they are infinitely easier to do that driverless cars, namely: viable alternatives to driving! Trains, streetcars, bike lanes, whatever.
Holding executives responsible for actual violence is considered promoting violence on this site and is not allowed. Cue the handwringing and moralizing from the usual suspects.
I think jail time for executives should be table stakes. Another thing would be fines well in excess of $10 million. The fines should be defined as percentages of gross revenue, or maybe even (to target VC-funded operations that operate at a loss) percentages of gross expenses. The penalties should be such that a few crashes can put the company entirely out of business.
Then it’s an okay cost of doing business. $10 million is a lot of money and consequences for these companies are not purely legal they are also social consequences.
Seems good. I'm a big Waymo user (344 rides) and love it, but I think they violate both traffic laws and common-sense courtesies of traffic in ways not captured by safety / crash statistics. Tickets probably are a great signal for ways the model needs to be improved.
For example, every time a Waymo picks me up from my apartment, it blocks a full lane of traffic on an extremely busy street, rather than pulling into a much quieter side street that an Uber driver will always use. I suspect (but have no idea), a lot of these low-level annoyances might be invisible to someone only looking at aggregated crash statistics, ride times, etc.
In many ways, I suspect the AI future might be better in many of the ways we can measure, but worse in those which aren't legible to statistics.
Don't they have 360 video of everything? Maybe they're low priority issues for now but surely those issues cannot be "invisible" because they "aren't legible to statistics."
There's no external public visibility into it, so other than wilbeddow writing on HN, how would anyone else know about this issue? I have the opposite problem, where the Waymo takes the legal option on an unbusy street, making the route a lot longer, rather than making an illegal U-turn that a human driver would do, when there's zero traffic or pedestrians it could remotely possibly run into.
I've actually thought of a much more dystopian idea: that Waymos could be technically used as roving traffic cameras, and report on the human drivers around them. They absolutely have strong enough telemetry systems to be able to determine things like excessive speeding, dangerous lane changes, red light running, etc., and their imaging systems could probably pick up a license plate with little additional modification... it's obviously not great from the perspective of general optics and morale, but it would surprise me if no one had floated basically WayNarc as a business model...
I do not live in California and am not up to speed on this issue, but as a casual observer I am shocked that they were allowed on the road without being ticketed for violations like any other vehicle operator.
Ticketing is a weird thing to do with driverless cars.
If the violations are intentional and easily fixable, then just pass laws/regulations requiring AV's to follow rules or else cease operations entirely.
If the violations are unintentional but happen only rarely in weird edge-case situations, then just set low frequency thresholds for them to be allowed, the same way we allow tiny amounts of rodent hairs in peanut butter. If AV companies exceed the threshold, then they get fined at first and eventually lose their permit -- but these aren't tickets for individual violations, but rather a yearly fine for going above the yearly threshold.
If the violations are intentional but not easily fixable -- e.g. they're stopping where not allowed because there's no legal place to stop within 15 blocks -- then the laws/regulations are bad, and tickets are essentially an unfair tax. That's the case in my city where moving trucks are essentially illegal, because it's illegal to double-park them, but there's usually no legal parking available within any reasonable distance that movers could carry furniture. So you just know that the cost of moving includes a "tax" of a parking ticket, unfair as it is.
Finally, if the violations are unintentional but happen all the time, the AV company should lose its permit because its software sucks.
I don't see how ticketing AVs for individual violations makes any sense.
EDIT: for those who think I'm letting AV companies get off too easily, it's precisely the opposite. I'm saying that if AV companies are violating traffic rules all the time and can't fix it, they should be banned. Ticketing is not the answer, because ticketing isn't holding these vehicles to a high enough standard. It's letting the companies get off the hook by merely paying occasional tickets instead of improving their software.
In all of your situations except for cases where no good legal option exists, ticketing is just the easier way to apply your suggested idea. It gives a direct incentive to the company to lower the rate as far as is possible. It doesn't allow some minimal amount without a fee, but that doesn't seem like that big of a deal.
The biggest reason for the difference between Autonomous vehicles and peanut butter is that with autonomous vehicles, we already have a compliance system in place....cops. It's not designed for autonomous vehicles, and you are correct that it's not the way you would design it for the ground up for autonomous vehicles, but it's far better to accept the imperfections than to build some new, separate compliance and monitoring system on top of the existing one. The benefits aren't large enough to justify it.
In the far future when the vast majority of vehicles are autonomous? Sure, probably worth scrapping to a new system (by then, my guess is that issues are rare enough to just not have a system at all and just use the legal system in the rare cases of large issues).
Until then, ticketing in the case of traffic violations seems fine and good enough to me.
You make some good points, but here are some counterpoints:
There is an existing infrastructure for ticketing by license plate, payment processing, collection, etc.
You’re describing changes to the law, which require a bunch of procedural hurdles. It’s much easier for the DMV to just promulgate new rules that tap into existing infrastructure, as they did here.
Also, how is the government supposed to assess whether these violations are intentional or not? Tickets are strict liability (you get the ticket if you do it regardless of intent, reasons, etc.) because it is easy to administer.
Of course I'm describing changes to the law. AV's inherently require tons of changes to the law. They already have. Permits for AV companies operate under new law. That is not an obstacle.
No, I think ticketing is the right thing to do. You set a law. Any instance of breaking that law costs money, so the AV company has an incentive to reduce the number of violations. The won't be able to bring the number of violations down to 0 just like we can't bring the number of cockroaches in chocolate down to 0, but that nonzero amount is just a regulatory cost they can decrease by getting closer to the goal of 0 violations.
Obviously, we should also have the option to pull vehicles that are brazenly ignoring the law and just eating the cost of the tickets. Just like we do with drivers who do that. But that should be the second line of defense if regular monetary fines (tickets) fail
The point is, with software you don't need tickets. Either the software is written to try to follow the law or it isn't. If it's trying, then we establish thresholds. If the company is actively trying to break the law, it should be shut down.
Tickets are a silly, roundabout way to go about it. They make sense for human drivers because they're all running different independent "brain software" and it's unrealistic for minor violations to ban someone from driving. But with shared software across a fleet, you can just require the company to fix its driving software directly when possible. Ticketing is actually counterproductive, because it allows these companies to avoid many of these fixes if the tickets are infrequent enough.
Ticketing AVs for individual violations like human drivers is the only fair way.
How would your proposal work for personal driverless cars, with/without custom modifications? ie. if my personal car commits violation on its way to pick me up
If you purchase an AV car then similarly it's up to the state to regulate the manufacturer. How could you possibly be personally responsible for the fact that it ran a red light?
And nobody should ever be allowed to personally modify an AV's software. Such a vehicle should never be allowed on the road.
Ticketing in California generally results in revenue going directly to the enforcing locality, not the state. It's an important difference, and why you tend to get things like speed traps for passing motorists
>If the violations are intentional and easily fixable, then just pass laws/regulations requiring AV's to follow the law or else cease operations entirely.
I have to stop reading the rest of the comment right there.
If the violations are intentional and easily fixable is an incredibly loaded presumption to start any type of conversation, dialogue, or debate. To the point, asking the question 'how do we qualify intention? How are we measuring difficulty of fix? Costs of payroll, computer, deployment, and potential regression testing? What about the very nature of the context that led up to it? Did an external 3rd party cutoff a robotaxi and require that the robotaxi veer into the oncoming traffic lane, bc sensors indicated it was the best decision to avoid a collision, prioritizing safety and human life over traffic law?
What happens when following traffic law statistically leads to a greater risk of loss of life over violating the law?
I must insist we move the dialogue upstream to reality as-is, and there is plenty to discuss there.
I will in good faith issue a starting point: how should we measure the robotaxi driver license wrt suspension? Do we issue a point system that is averaged across the fleet, e.g. violations/car before suspending all operations until licensure evaluation? Personally I think that is a fair starting point amd am completely open minded to alternative views.
Are they trying to drive safety or revenue? The second order effect people forget about is tickets are a source of revenue for cities and police depts. Surely driverless car companies will absorb a few tickets and fix the issue quickly.
So I do wonder what happens in the future when roads and cars are all automated and city funding from this channel dries up.
>ever stricter rules that cause tickets despite technology getting better.
That sort of stuff might work on a bunch of peasants since you can screw them individually for small enough amounts it doesn't meet the threshold for political pushback but it won't fly against a few megacorps who do self driving fleets.
I imagine the city funding issue could be solved with some sort of tax to operate within the city, where a couple cents from each mile driven would be paid to the city. Alternatively, a higher cost for registration at the state level.
What I worry more about is a future where private car ownership seems impractical when there is a large fleet of autonomous Ubers out there to handle the day-to-days, which start out cheap. Once society reaches a point of dependence, will there be enough competition to keep the price down, or will we see consumers of the services get squeezed as companies ratchet up prices to increase margins.
Given the lack of enforcement it must not be true that it’s a source of revenue for the city. I see ticketable violations multiple times a day and zero enforcement
Fix the issue quickly, or optimize to the point where revenue gained from breaking the law exceeds the fine. Last I read they were holding steady on "passengers want us to pull into bikelanes to drop-off" in California.
At the same time, there are not many cops making USD200K per annum in my municipality. And it's in flyover country, where everyone is clamoring for lower taxes. So I think it'd be a bit naive to think politicians wanting to score easy points with voters in cities, and even states, won't take the opportunity to extract a bit of revenue out of Big Tech.
Not saying it's right. Just saying that's how local politics work.
Bike Lanes have turned out to be an interesting edge case.
Waymos are currently dropping off and picking up passengers in a bike lane which is not legal (because it is dangerous) however many ride share drivers also do this. As somebody who is commonly a biker / pedestrian I am excited that AVs will likely make many things safer for that class of user. That being said, I do worry about how we encode these "social understandings" of laws.
- A waymo I rode in on a highway was happy to go slightly above the speed limit
- It seems at stop signs waymo prefers to be slightly aggressive to make it through rather than follow the letter of the law.
It seems silly that we have to teach robots to break certain laws sometimes but parking in bike lanes / yielding to pedestrians are laws that human drivers break all the time and I hope the mechanisms mentioned in the article prevent us from teaching robots to program anti-social but common behavior.
As a cyclist and driver I figure you have to use some common sense. I probably break some regulations all the time like stopping where you are not supposed to briefly but being safe and not inconveniencing others is the main thing.
It's all pretty nuanced. I don't know where to draw a line.
For instance: Busy intersections with 4-way stop signs are an interesting example of how laws don't quite fit.
It's obviously important to get the order right since nobody wants to be in a car crash today. But the law (often -- we've got 50 states worth of driving laws and they aren't all the same) says something very specific and simplistic about the order: First-come, first-served; if order is unclear, yield to the right. Always wait for the intersection to be completely clear before proceeding.
That sounds nice and neat and it looks good on paper. It was surely at least a very easy system to describe and then write down.
But reality is very different: 4 way stops are an elaborate dance of drivers executing moves simultaneously and without conflict. For instance: Two opposite, straight-going cars can proceed concurrently works fine. All 4 directions can turn right, concurrently. Opposing left turns at the same time? Sure! While others are also turning right? Why not.
When there's room for a move and it creates no conflict, then that move works fine.
We all were taught how these intersections are supposed to work, but then reality ultimately shows us how they do work. And the dance works. It's efficient. Nobody gets ticketed for safely dancing that dance. (And broadly-speaking, a timid law-abiding driver who doesn't know the dance will be let through...eventually.)
The main problem with the dance is that it's difficult to adequately describe and write down and thus codify in law.
The nuance for four-way stops is pretty simple. First come, first serve queue. Except you are allowed to jump out of order if you jumping out of order doesn't slow the people ahead of you down.
I read an article a while back that they made Waymo more aggressive, in the ways you mention, because they were quite annoying to other drivers when following the letter of the law. There is something to be said for following the flow of traffic.
I would imagine they would be able to revert back to more strict rule following once autonomous vehicles reach some level of critical mass and human drivers are needing to adapt to the AV traffic, rather than AVs needing to adapt to human traffic.
This is false. It is only legal in the rare event that a passenger requires curb-side access for accessibility/ADA reasons; any other use is still illegal. To quote SFMTA taxi training:
Only drop off in a separated bike lane if you
have disabled or elderly customers who
require direct access to the curb
You may only pick up in a separated bike lane
if the dispatcher tells you that the customer is
disabled and must be picked up at a location
that is next to a separated bike lane.
Taxi drivers often intentionally misstate this regulation because it’s more annoying to follow the law and find a legal place to stop so they pretend they are allowed to use bike lanes for any reason.
Yea, California civic code is pretty liberal with curbside parking parking, allowing it anywhere it isn't expressly prohibited, with signs declaring it so.
I live in Northern California, inland of San Francisco, and the city closest to me has a bunch of streets with bike lanes that are just painted onto the shoulder and otherwise are legally just a shoulder. Most of those streets also prohibit parking, but some don't, so parking is in the bike lane.
It gets really crazy in the denser parts of Southern California, where parking is sometimes not prohibited even when there isn't a shoulder, so parked cars full-on block a driving lane.
Good luck. This is going to be faceless machine v. faceless machine. I can't think of anything more useless in terms of fixing actual needs of society. In fact, I expect it to lead to shady business. Example: Waymos are all paid up at this intersection so traffic is allowed to safely ignore the rules for this month. Tesla Cybercab is still fighting us, so we continue to ticket them.
What's gonna really be funny is the first time a state legislates that an AV company has to keep a bug in their software to maintain a municipal income flow.
As a Waymo (and other driverless car) supporter, this seems like an obviously good thing, right? I’m a little surprised this wasn’t possible before given the amount of regulatory scrutiny (correctly) applied to these companies.
> As a Waymo (and other driverless car) supporter, this seems like an obviously good thing, right? I’m a little surprised this wasn’t possible before given the amount of regulatory scrutiny (correctly) applied to these companies
Not necessarily. I went into a bit more detail in my own comment but it might be useful to think that when regulations are written keeping in mind multibillion dollar automobile companies, what the effect of those regulations on a person maintaining their own vehicles might be.
Consider that your Waymo got ticketed, but you had flashed it with a "no customer telemetry" firmware. Once Waymo gets the ticket, they flag your car as having "unauthorized" software and now the ball's in your court that the reason why your Waymo got ticketed has nothing to do with the telemetry feature that tells Waymo what radio stations you were listening to.
Also, when regulations are written keeping in mind multibillion dollar automobile companies, the ticket isn't going to cost $500.
They should automate the fines by requiring the cars to send the video for processing through an enforcement program. Self-driving, self-ticketing. Maybe that would help force manufacturers to improve quality and not hide as much.
Laws should be loser for autonomous vehicles with good safety records.
No one is protected by preventing waymos from making rolling stops, and driving like a human Uber driver.
Ideally the fees would be similar to the Norway model, where some tickets are tied to the income of the driver, in this case the pre-tax earnings of the company that created the driverless car.
That can make sense (opinions differ) for individuals, but it's not like the company is advertising with "we get you there at 1.2x legal speed". They're not competing on that; they're not choosing to do this on purpose like an individual might choose to speed (for example because of economic incentives if their hourly price is high)
If they were, then it makes sense to fine them to some multiple of the benefit they got from this advertising tactic, but as it is, I don't see why it should be different from anyone else's ticket. The company isn't likely to enjoy a flood of this administrative work, besides the cost of the actual fines, so they'll work to minimise them anyway
They may not advertise “getting you there at 1.2x legal speed” but the sooner they drop you off, the sooner they can get another fare. Over a whole fleet it will add up to changing the size of the required fleet.
If getting a ticket one ride in a thousand is cheaper than deploying another 2000 cars to make up for the increased trip time I’d expect them to keep getting tickets.
I’m also not sure they don’t do it on purpose. Tesla self driving has an aggressive mode willing to speed and roll through stop signs. Those were deliberate, law breaking, choices.
Assuming you divide it down to the earnings per car, that makes perfect sense. Of course right now they aren't making any profit at all, and by the time it is relevant it is likely that the cars will commit substantially no violations at all.
Isn't Norway only for drunk driving? Finland has it for massive speed excesses, but it is based on net taxable income taking out business expenses for taxi drivers, and Waymo is still negative.
If they become profitable you'd want to normalize by number of miles, unless you just want an incentive system to get more people on the road (extra drivers) and increase chance of humans suffering road injuries to boost employment in an internal service sector.
But even then coming out with a more efficient fleet than a competitor for higher margin would be penalized. You'd rather disincentivize skimping on safety for margin and not disincentivize better maintenance and fuel economy.
Dunno what it's like in Cali but in the UK nearly all tickets are generated by automatic cameras and arrive in the post to the registered keeper of the vehicle.
This will be just another minor cost of doing business unless they are treated like human drivers in at least two other ways.
1) If theses companies get enough points on their license, their license is revoked. Not just for that vehicle, but for all of their vehicles. (The number of points would need to be adjusted for number of miles driven.)
2) Senior executives could be held criminally liable for vehicular manslaughter the way a normal drivers are. A death doesn't mean someone is going to prison, but their would be a police investigation. If an exec decided to ship a product with a known bug that lead to someone's death it should be treated with the same seriousness as a drunk driver killing someone.
If there aren't serious consequences for driverless cars committing crimes (I mean jail time for executives serious), what's to stop someone for starting a hitman business?
We'll just run someone over with our "driverless" car and pay a fine - capitalism, baby!
UPDATE (can't respond to the two subcomments below due to post throttling, so I'm updating this comment instead)
> the car is basically a taxi and the taxi service is to blame for any mistakes
@skybrian - Agreed! but if you read the article, the CA DMV is ticketing the manufacturer, not the operator.
None of my concerns hold if the operator was ticketed - infact, existing regulations are set up exactly that way, so no new regulation was even necessary. Something's not adding up.
> Right now, no one can independently own and operate an AV the way Waymo or Tesla does
@ourspacetabs - Sure but the regulation seems to be specifically addressed at the manufacturer, not the operator.
I would have no concern if the regulation was addressed to the operator. The article atleast doesn't imply that's the case.
---
> The state's Department of Motor Vehicles (DMV) has announced new regulations on autonomous vehicles (AVs), including a process for police to issue a "notice of AV noncompliance" directly to the car's manufacturer.
> Under the new rules, police can cite AV companies when their vehicles commit moving violations. The rules will also require the companies to respond to calls from police and other emergency officials within 30 seconds, and will issue penalties if their vehicles enter active emergency zones.
These are new frontiers in automotive regulation. Typically, if a car failed because of a manufacturer issue, the driver would be ticketed. For example: if Hyundai sold vehicles where the engine would explode around 50k miles and that caused an accident, the driver of the vehicle would be ticketed for it.
Now if we take the human out of it, it is Hyundai that would be ticketed for it. Insurance companies are certainly going to take notice and adjust their risk models accordingly.
I imagine there will be a lot of fingerpointing by the manufacturer towards customers.
In the worst case, this is the end of customers servicing their own autonomous vehicles.
If we imagine that most vehicles in the next 15 years will be autonomous, this would mean customers would have to handle regulation aimed at multibillion dollar companies, if they were to service their own autonomous vehicles, or give up on servicing their own autonomous vehicles entirely and just rent them instead.
Not sure I agree. The clear boundary here to me is who owns and is operating the vehicle. Waymo both owns and operates their vehicles, it’s a taxi service, you wouldn’t say a Waymo rider is operating a vehicle and therefore deserves the ticket. Right now, no one can independently own and operate an AV the way Waymo or Tesla does.
When that happens someday, then the ticket would go to the owner/operator of the vehicle - whoever bought the car. If you get a ticket due to something dumb your personally owned Waymo did, wouldn’t you pursue that case against Waymo separately, the same way you’d pursue Hyundai for selling you a car whose engine blew up after 50k miles?
It seems pretty reasonable to me that when you're not driving, the car is basically a taxi and the taxi service is to blame for any mistakes. The car manufacturer isn't just making cars anymore. It's providing a service.
Perhaps they could sell the car to a different taxi service, though?
maybe Tesla can put that weird robot that connects the charging connector to the car to use by building a robot that can give the police a hand to place the ticket into
I don’t disagree with needing some sort of consequence for bad driverless actions. But I distrust the motivation. Maybe California is just looking for more revenue sources after rampantly mismanaging their state and letting corruption and fraud continue.
I am, in general, hoping AV will reduce road deaths in the future.
The last hurdle is regulatory. We can’t let AV manufacturers use “there’s no driver” as a way to escape responsibility, externalizing the harms AC cause onto society.
The question is how to achieve fairness. If a human driver commits vehicular manslaughter, they get the book. What about AV? $10 million? Executives go to jail? What if $10 million fine per X AV miles driven is an OK cost of doing business?
> If a human driver commits vehicular manslaughter, they get the book.
Hah. Do they, though? https://sfstandard.com/2026/03/20/mary-lau-sentenced-probati...
The standard for human drivers is through the floor.
The reason that’s a news story is because the outcome is unusual.
When things are normal and happening all the time, they’re not reported as abnormal outcomes.
The world is a big place. Being able to think of a counter-example does not negate a general point.
21 replies →
Freakonomics did a pod about this, titled “how to get away with murder”.
see https://sf.streetsblog.org/2026/03/06/motorist-careens-onto-... and see what the police said to the driver…
Who does it benefit if an accident ruins a second life?
What does a jail sentence deter? ("[no] gross negligence [...] wasn’t engaging in a race or sideshow, was not texting, and was not under influence")
This person was 80 years old with no criminal record, needs to pay $67400 in restitution, do 200 hours of community service, isn't allowed to drive for 3 years but "never intends to drive again". Apologised to the family of the victims. She's taking responsibility and I can't imagine forced labor at that age is fun. What more can you ask for here? The family member isn't coming back if she gets what's not unlikely to be a life sentence
Edit:
> She told a witness at the scene that she was trying to park her car when she accidentally moved her foot to the gas pedal.
This seems to happen a lot. Don't know about statistics but this happened to someone I know at 50yo (thankfully only damaged their own car minorly), and you hear it on the news with some regularity. Maybe the gas needs to be in a fundamentally different spot from the brake? We can jail the people to whom it happens, sure, but I can understand a judge using their head instead of their heart. The real solution must come either from the automotive industry or legislation
45 replies →
Better than the current standard for AV, which is "what floor?"
9 replies →
> The standard for human drivers is through the floor.
The linked article doesn't describe the standard. It describes a single, exceptional example.
20 replies →
In the US, 11 deaths per billion miles driven (or about 47k per year) is currently seen as an OK cost.
More than twice as much per mile as places like Sweden and Switzerland, and still substantially more than places like Canada, Australia or Germany (all three in the 6-8 deaths per billion miles range). So it's not like there isn't room to improve. The effort to do so just isn't seen as worth the cost at the societal or government level
Turning that into a monetary cost would change the ethics slightly, but it wouldn't be a monumental shift
The issue here is that a lot of the concerns about AV's are orthogonal to the standard metrics of concern.
I'm a strong transit alternatives advocate, but even I recognize that a firetruck or ambulance being blocked by an AV has the potential to cause an outsized amount of death and destruction, because deaths aren't always linear and a fire that is able to get out of control can do catastrophic damage compared to a single out of control vehicle.
I'm genuinely stunned that AV's do not have the ability to be "commandeered" by Police/Fire/EMS in a pinch, and I'm honestly surprised that regular citizens can't just hit a red button that signal "this is seriously an emergency." These are fairly simple steps to mitigate the tail risk of AV's but the platforms aren't going to prioritize that if there are no incentives.
6 replies →
Don't forget to add rail incidents to that metric. I live in Spain, this year we had 4 derailments for a total of 48 deaths and 195 injured. The USA has had 0 passengers killed or injured from train accidents this year. Portugal had 15 death after a tram derailment. In Amsterdam, the tram is more dangerous than the car.
Also Germany is very high (for European standards) because of the Autobahn. They can save around 140 lives a year by having a limit on the Autobahn but the car lobby in Germany is very strong. Those 140 lives are seen as an OK cost just to go vroom on the Autobahn.
13 replies →
Coming from a bio background, I’ve always been confused why auto fatality stats are normalized per miles driven. Epidemiological metrics like incidence or prevalence seem like they would work fine? Town A would be “safer” than town B if people’s commutes are 20% shorter, even if accidents occur w same frequency
6 replies →
> it's not like there isn't room to improve
Losing one's license means destitution for many Americans. That places practical limits on enforcement compared with less car-oriented countries.
28 replies →
> So it's not like there isn't room to improve. The effort to do so just isn't seen as worth the cost at the societal or government level
That effort being what, exactly?
Road fatalities per mile driven don’t translate cleanly from country to country because the type of roads and even types of deaths (single vehicle, multi vehicle) are different.
We could set the speed limit at 25mph everywhere and force all vehicles to not exceed that limit and that would make the number go down, but the cost would be extreme for everyone.
So what, exactly, are the solutions you are proposing?
1 reply →
> 11 deaths per billion miles driven
You should calculate how many are "single vehicle accidents" and how many are "multiple vehicle accidents." In the US the majority are single vehicle.
> seen as an OK cost.
You cannot build a system that stops every stupid person from doing something stupid without introducing absolute tyranny.
1 reply →
Doesn’t that 11 per billion statistic include commercial drivers as well? And doesn’t the United States have by far the largest percentage of commercial miles driven of any developed nation?
There’s a far cheaper solution available. Log books.
We subsidize driving by somewhat over a trillion dollars annually, mostly due to lax penalties for negligence which shift liability to drivers’ victims[1]. One way to tackle all of these problems would be requiring drivers to cover the full damages.
Another simple and effective measure would be changing fines from absolute values to a percentage of income. Right now, parking in a bike lane usually doesn’t kill anyone so drivers are only thinking there’s a small chance of a small fine, but if it was a chance of, say, 0.1% of annual income Waymo technology would magically be capable of not doing that. Add a right of private action and enforcement would be high enough to really speed things along, too, and that’d improve safety and travel times for all road users.
1. https://www.economist.com/united-states/2024/01/18/why-car-i...
Yeah, making fines relative to income would change behaviors for sure. A $20 ticket when you make $20 an hour hits different when you're making $200 or $2,000/hr. If it was a percentage of pay, then the ticket would actually sting.
1 reply →
> If a human driver commits vehicular manslaughter, they get the book.
If only! "10 Days In Jail For Drunken Driver Who Killed Cyclist Bobby Cann" https://www.dnainfo.com/chicago/20170126/old-town/ryne-san-h...
I almost feel bad for noticing this, but:
> San Hamel was a partner in a business called AllYouCanDrink.com at the time.
…
> Cann, an experienced cyclist who once biked from New Hampshire to Chicago, was heading home from his job at Groupon the night he was killed.
It looks like allyoucandrink.com now redirects to Groupon, in a decent bit of irony.
> We can’t let AV manufacturers use “there’s no driver” as a way to escape responsibility, externalizing the harms AC cause onto society.
There is essentially nothing to be gained from doing this because it will not in either case be manufacturer; it will be an insurance company.
If the liability is paid by the vehicle owner's insurance then things work as they do now. You buy a car, insure it, if there is a liability there is an insurance claim and then the victim has someone to pay them for their injuries. Meanwhile the manufacturers still have a financial incentive to make safer cars because buyers want neither accident prone vehicles as the one they use nor high insurance rates. The insurance rates in particular are in direct competition with the car payment for the customer's available income.
Whereas if you try to put the liability on the manufacturer, several stupider things happen.
First, they're just going to buy insurance anyway, but now the insurance cost has to be front-loaded into the purchase price, which increases costs because now you're paying car loan interest on money to cover insurance five and ten years from now, when you otherwise wouldn't have needed to pay the premiums until the time comes.
Second, what happens to cars from manufacturers who no longer exist? They can't continue paying for insurance if they're bankrupt, so you need it to be someone else. Worse, if a company produces a vehicle which is unsafe, that will tend to cause them to go bankrupt. But then people still have them, and would continue to operate them if they're allowed to point victims at the bankrupt manufacturer, whereas the incentive you want is for the premiums on those cars to go up for the vehicle owners so that they stop operating them.
> If a human driver commits vehicular manslaughter, they get the book. What about AV?
They get their licenses pulled statewide [1]. Cruise's single negligent manslaughter event carried more consequence than dozens of human cases combined.
[1] https://www.dmv.ca.gov/portal/news-and-media/dmv-statement-o...
I also hope AV will reduce road deaths in the future but I don't think what will make the difference is regulatory. Rather the tech will advance from doesn't work to works in Waymos but is expensive to works in most cars and has become cheap.
> If a human driver commits vehicular manslaughter, they get the book.
I wish this were true. Often they get off with a light punishment, or no punishment at all.
> What if $10 million fine per X AV miles driven is an OK cost of doing business?
It’s the same cost/benefit we accept under current rules. Why have cars that can go 3x the speed limit? Why not require breathalyzers in cars before starting them? Why not fine logistics companies if one of their drivers breaks the law? And so on… Because it’s worth it
>Why not require breathalyzers in cars before starting them?
FYI Cars will soon detect if you are impaired.
Your questions are pertinent but what’s the benefit "worth" you’re referring to? The two first proposals would risk a politician popularity and the last one would be lobbied to he’ll buy the logistic companies. IMHO inconvenience isn’t worth driving among drunk coursier at 200kmh.
What happens if you build a bridge and it breaks?
These people want to play god with our lives but at the same time move fast and break things. Look at software quality anywhere, it's a mess and only about to get much worse.
We should not let them. Jail time for anyone involved in any of the decision making process, applied at scale with the number of vehicles and deaths.
Why should the standards be any different? They want to change the status quo with tech only so they can get paid and extort us with yet more subscriptions.
AVs will never substantially reduce road deaths. They will optimize to just being slightly better than human, but fail in new and more unexpected ways. There is not enough incentive for them to make it safer.
If we consider fairness/retribution/justice then we won't get this future of less road deaths.
1. There will always be a probability of death from a vehicle. This can never go to 0%.
2. If the probability of a AV causing death is many magnitudes lower than human driving then that is the future we must choose.
If 1 and 2 holds and we hold AV manufactures accountable in the sense that Executives go to jail or are personal liable financially for deaths/injuries then AV will never get released or become mainstream even if this results in less total deaths. The sense of fairness/justice/retribution may make us feel better but result in more overall deaths. Logically this means that there must be a standard. Something like x deaths per y cars manufactured. If above the threshold you get big fines as a company. As technology gets better you can lower the threshold. Anything apart from causing deaths either purposefully or negligently would have be ignored.
Can we as a species accept this? That is another question.
We can look to other forms of automation to get a sense of what to do. For example, planes largely fly themselves and a loss of life due to manufacturing errors from the manufacturer would deem them liable for those deaths. Seems like the solution here is large penalties and generally broad disincentives for incurring harm.
> What about AV? $10 million? Executives go to jail? What if $10 million fine per X AV miles driven is an OK cost of doing business?
If AVs will save lives, we need to be sure we aren't punitive to the point of making them disappear.
Adjust the fines such that X is some acceptably large number.
The trickiest part will be figuring out how many dollars per mile driven is an acceptable cost of business..
I'd probably reserve the whole executives to jail thing to cases where you can prove negligence or something.
Societies can already reduce road deaths to nearly zero, it's cheap, it's easy, and it's fun. It's just redirecting all of the cash we spend on vehicles/cars/highways/roadways/signs/etc into public infrastructure that is all encompassing.
A hundred billion dollars a year [0] on construction (reading the definition I'm not 100% sure what is included in this due to how definitions can be hazy) has goes a long way, not to mention the amount we spend on gasoline, car maintenance, etc etc.
The reason I say it's fun, is because I love being on a train. First time I was able to ride one, which due to living in the good old USA wasn't until I was 23, I yelled "I'm on a train" . The Germans traveling with me weren't as into it.
[0] https://fred.stlouisfed.org/series/TLHWYCONS#
Just because you like trains does not mean that it is actually a solution to everyone’s problem. For example, until proper law-enforcement starts happening on public transit, nobody in my family is allowed to take it in USA (they are allowed to do so in Singapore or Japan)
4 replies →
Full liability. It's a machine with predictable performance.
The law applied to humans needs to account for their fallibilities. Not so with a machine.
The CEO gets charged with manslaughter? I work in healthtech and the responsible individual is certainly personally liable for any harm that results from reckless behavior, it should be the same here.
Same as if someone were driving, if a person just jumps in front of your car while you're driving under the limit/sober/etc, you aren't at fault, so the AV should also not be at fault if it couldn't reasonably avoid the harm. You balance these things, benefit to society vs harm to society, and you come to an acceptable tradeoff.
Could you provide examples of healthcare executives held personally liable for harm resulting from reckless decision-making? I have never heard of such a thing happening in healthcare so framing CEO responsibility as a solution to the problem sounds like a stretch to me.
Some examples: Elizabeth Holmes got canned for lying to investors, not harming patients. Purdue Pharma plead guilty to misleading regulators and giving doctors kickbacks, not causing some hundreds of thousands of opioid deaths, but no Sackler family members were personally tried.
1 reply →
> The CEO gets charged with manslaughter?
Well then forget autonomous vehicles altogether and allow the human joyride to continue, because no CEO is stupid enough to risk that.
> The CEO gets charged with manslaughter? I work in healthtech and the responsible individual is certainly personally liable for any harm that results from reckless behavior, it should be the same here.
This is in like China, yes? Certainly not in the US of A, hence Luigi and all that…
Want to reduce road deaths? Invest in public transportation.
We've had public transportation for a couple of centuries and no where has it really led to a road death free utopia. I like public transport but no harm trying something new.
I had to look up a name for this. "Utopian Fallacy."
You don't have to get rid of genuine progress just because your utopian vision has something better. The USA is on the path to autonomous vehicles. They are not on the path to public transportation excellence.
Yup. Even if "safer per mile", more cars and more miles driven will probably outweigh the benefits. And still be hazardous to cyclists and pedestrians, still make us design stupid cities (built for cars, not people), etc.
Like how electric cars were for saving the car companies, not the planet, autonomous will be the same.
Simple. Blame the owner of the vehicle. They relied on automation and it failed. They go to jail for negligent homicide (whatever flavor is appropriate). That will tank sales of any AV tech that cannot maintain standards.
People are killed by industrial equipment fairly regularly.
I'd say we actually have a perfectly functional legal framework for all of this, and the real issue is a lot of new people are about to find out it also applies to them as well.
Whether it was working well in the first place is the real question.
Now try applying this logic to elevators.
Many things already reduce road deaths and they are infinitely easier to do that driverless cars, namely: viable alternatives to driving! Trains, streetcars, bike lanes, whatever.
The legal entity driving the AV should of course be responsible in the same way as human drivers are.
My understanding is that that is already the legal situation?
Holding executives responsible for actual violence is considered promoting violence on this site and is not allowed. Cue the handwringing and moralizing from the usual suspects.
> The last hurdle is regulatory
How’d you arrive at this conclusion? Why would fleet providers accept regrettable losses? Wouldn’t the last hurdle be technical?
> The question is how to achieve fairness
What does that have to do with automotive safety?
I think jail time for executives should be table stakes. Another thing would be fines well in excess of $10 million. The fines should be defined as percentages of gross revenue, or maybe even (to target VC-funded operations that operate at a loss) percentages of gross expenses. The penalties should be such that a few crashes can put the company entirely out of business.
> hoping AV will reduce road deaths in the future.
It won't. The majority of fatalities are caused by drugs and alcohol.
> The last hurdle is regulatory.
Indeed. Compare the USAs DUI laws with any other first world country.
Then it’s an okay cost of doing business. $10 million is a lot of money and consequences for these companies are not purely legal they are also social consequences.
Seems good. I'm a big Waymo user (344 rides) and love it, but I think they violate both traffic laws and common-sense courtesies of traffic in ways not captured by safety / crash statistics. Tickets probably are a great signal for ways the model needs to be improved.
For example, every time a Waymo picks me up from my apartment, it blocks a full lane of traffic on an extremely busy street, rather than pulling into a much quieter side street that an Uber driver will always use. I suspect (but have no idea), a lot of these low-level annoyances might be invisible to someone only looking at aggregated crash statistics, ride times, etc.
In many ways, I suspect the AI future might be better in many of the ways we can measure, but worse in those which aren't legible to statistics.
Don't they have 360 video of everything? Maybe they're low priority issues for now but surely those issues cannot be "invisible" because they "aren't legible to statistics."
There's no external public visibility into it, so other than wilbeddow writing on HN, how would anyone else know about this issue? I have the opposite problem, where the Waymo takes the legal option on an unbusy street, making the route a lot longer, rather than making an illegal U-turn that a human driver would do, when there's zero traffic or pedestrians it could remotely possibly run into.
Which is also not captured in the statistics.
I've actually thought of a much more dystopian idea: that Waymos could be technically used as roving traffic cameras, and report on the human drivers around them. They absolutely have strong enough telemetry systems to be able to determine things like excessive speeding, dangerous lane changes, red light running, etc., and their imaging systems could probably pick up a license plate with little additional modification... it's obviously not great from the perspective of general optics and morale, but it would surprise me if no one had floated basically WayNarc as a business model...
1 reply →
I do not live in California and am not up to speed on this issue, but as a casual observer I am shocked that they were allowed on the road without being ticketed for violations like any other vehicle operator.
Any other vehicle operator is also rarely ticketed in California.
Completely agree. How is it possible they issued these permits (years ago it seems) without having this infrastructure in place?
Scooter companies are allowed to violate ada putting the scooters all over the sidewalk where there isn’t 3ft of space to pass them.
Ticketing is a weird thing to do with driverless cars.
If the violations are intentional and easily fixable, then just pass laws/regulations requiring AV's to follow rules or else cease operations entirely.
If the violations are unintentional but happen only rarely in weird edge-case situations, then just set low frequency thresholds for them to be allowed, the same way we allow tiny amounts of rodent hairs in peanut butter. If AV companies exceed the threshold, then they get fined at first and eventually lose their permit -- but these aren't tickets for individual violations, but rather a yearly fine for going above the yearly threshold.
If the violations are intentional but not easily fixable -- e.g. they're stopping where not allowed because there's no legal place to stop within 15 blocks -- then the laws/regulations are bad, and tickets are essentially an unfair tax. That's the case in my city where moving trucks are essentially illegal, because it's illegal to double-park them, but there's usually no legal parking available within any reasonable distance that movers could carry furniture. So you just know that the cost of moving includes a "tax" of a parking ticket, unfair as it is.
Finally, if the violations are unintentional but happen all the time, the AV company should lose its permit because its software sucks.
I don't see how ticketing AVs for individual violations makes any sense.
EDIT: for those who think I'm letting AV companies get off too easily, it's precisely the opposite. I'm saying that if AV companies are violating traffic rules all the time and can't fix it, they should be banned. Ticketing is not the answer, because ticketing isn't holding these vehicles to a high enough standard. It's letting the companies get off the hook by merely paying occasional tickets instead of improving their software.
In all of your situations except for cases where no good legal option exists, ticketing is just the easier way to apply your suggested idea. It gives a direct incentive to the company to lower the rate as far as is possible. It doesn't allow some minimal amount without a fee, but that doesn't seem like that big of a deal.
The biggest reason for the difference between Autonomous vehicles and peanut butter is that with autonomous vehicles, we already have a compliance system in place....cops. It's not designed for autonomous vehicles, and you are correct that it's not the way you would design it for the ground up for autonomous vehicles, but it's far better to accept the imperfections than to build some new, separate compliance and monitoring system on top of the existing one. The benefits aren't large enough to justify it.
In the far future when the vast majority of vehicles are autonomous? Sure, probably worth scrapping to a new system (by then, my guess is that issues are rare enough to just not have a system at all and just use the legal system in the rare cases of large issues).
Until then, ticketing in the case of traffic violations seems fine and good enough to me.
At some point though those tickets need to actually hurt and no be just a cost of doing business.
After enough violations humans get their license taken away. What happens after autonomous vehicle get enough violations?
4 replies →
You make some good points, but here are some counterpoints:
There is an existing infrastructure for ticketing by license plate, payment processing, collection, etc.
You’re describing changes to the law, which require a bunch of procedural hurdles. It’s much easier for the DMV to just promulgate new rules that tap into existing infrastructure, as they did here.
Also, how is the government supposed to assess whether these violations are intentional or not? Tickets are strict liability (you get the ticket if you do it regardless of intent, reasons, etc.) because it is easy to administer.
Of course I'm describing changes to the law. AV's inherently require tons of changes to the law. They already have. Permits for AV companies operate under new law. That is not an obstacle.
No, I think ticketing is the right thing to do. You set a law. Any instance of breaking that law costs money, so the AV company has an incentive to reduce the number of violations. The won't be able to bring the number of violations down to 0 just like we can't bring the number of cockroaches in chocolate down to 0, but that nonzero amount is just a regulatory cost they can decrease by getting closer to the goal of 0 violations.
Obviously, we should also have the option to pull vehicles that are brazenly ignoring the law and just eating the cost of the tickets. Just like we do with drivers who do that. But that should be the second line of defense if regular monetary fines (tickets) fail
The point is, with software you don't need tickets. Either the software is written to try to follow the law or it isn't. If it's trying, then we establish thresholds. If the company is actively trying to break the law, it should be shut down.
Tickets are a silly, roundabout way to go about it. They make sense for human drivers because they're all running different independent "brain software" and it's unrealistic for minor violations to ban someone from driving. But with shared software across a fleet, you can just require the company to fix its driving software directly when possible. Ticketing is actually counterproductive, because it allows these companies to avoid many of these fixes if the tickets are infrequent enough.
5 replies →
Seems to me like ticketing is a really simple proxy for everything you’ve just described.
Why pass a thousand new laws when the existing laws have an enforcement mechanism?
Ticketing AVs for individual violations like human drivers is the only fair way.
How would your proposal work for personal driverless cars, with/without custom modifications? ie. if my personal car commits violation on its way to pick me up
I'm talking about AV fleets.
If you purchase an AV car then similarly it's up to the state to regulate the manufacturer. How could you possibly be personally responsible for the fact that it ran a red light?
And nobody should ever be allowed to personally modify an AV's software. Such a vehicle should never be allowed on the road.
Yes, I thought AV by design should not voilate traffic laws.
Ok, but why are AVs getting a break on the same tickets a human gets no "low frequency threshold" for them to be allowed.
If a AV runs a red light or a stop sign, it should be the same penalty, period.
If AV companies want to avoid the tickets, they can make their claimed superior drivers avoid violating the law.
No, you're missing the point.
If an AV is regularly running red lights or stop signs, it should be a much worse penalty. It shouldn't be permitted to operate at all.
It shouldn't just be given occasional tickets. Tickets are not the right enforcement mechanism.
2 replies →
They want to make money from the tickets
I think part of ticketing is the state makes money off of it. If they just shut these companies down no one benefits.
Ticketing in California generally results in revenue going directly to the enforcing locality, not the state. It's an important difference, and why you tend to get things like speed traps for passing motorists
Traffic has rules, you violate them you get a ticket
>If the violations are intentional and easily fixable, then just pass laws/regulations requiring AV's to follow the law or else cease operations entirely.
I have to stop reading the rest of the comment right there.
If the violations are intentional and easily fixable is an incredibly loaded presumption to start any type of conversation, dialogue, or debate. To the point, asking the question 'how do we qualify intention? How are we measuring difficulty of fix? Costs of payroll, computer, deployment, and potential regression testing? What about the very nature of the context that led up to it? Did an external 3rd party cutoff a robotaxi and require that the robotaxi veer into the oncoming traffic lane, bc sensors indicated it was the best decision to avoid a collision, prioritizing safety and human life over traffic law?
What happens when following traffic law statistically leads to a greater risk of loss of life over violating the law?
I must insist we move the dialogue upstream to reality as-is, and there is plenty to discuss there.
I will in good faith issue a starting point: how should we measure the robotaxi driver license wrt suspension? Do we issue a point system that is averaged across the fleet, e.g. violations/car before suspending all operations until licensure evaluation? Personally I think that is a fair starting point amd am completely open minded to alternative views.
"begin" you mean they haven't been doing that already? That seems wrong
Are they trying to drive safety or revenue? The second order effect people forget about is tickets are a source of revenue for cities and police depts. Surely driverless car companies will absorb a few tickets and fix the issue quickly.
So I do wonder what happens in the future when roads and cars are all automated and city funding from this channel dries up.
Tolls, revenue taxes, ever stricter rules that cause tickets despite technology getting better.
>ever stricter rules that cause tickets despite technology getting better.
That sort of stuff might work on a bunch of peasants since you can screw them individually for small enough amounts it doesn't meet the threshold for political pushback but it won't fly against a few megacorps who do self driving fleets.
Police departments have already moved on from traffic enforcement to civil forfeiture. Like, a decade ago.
I imagine the city funding issue could be solved with some sort of tax to operate within the city, where a couple cents from each mile driven would be paid to the city. Alternatively, a higher cost for registration at the state level.
What I worry more about is a future where private car ownership seems impractical when there is a large fleet of autonomous Ubers out there to handle the day-to-days, which start out cheap. Once society reaches a point of dependence, will there be enough competition to keep the price down, or will we see consumers of the services get squeezed as companies ratchet up prices to increase margins.
Given the lack of enforcement it must not be true that it’s a source of revenue for the city. I see ticketable violations multiple times a day and zero enforcement
Fix the issue quickly, or optimize to the point where revenue gained from breaking the law exceeds the fine. Last I read they were holding steady on "passengers want us to pull into bikelanes to drop-off" in California.
Probably higher city/state taxes. A police officer making over $200k a year with a pension isn’t making most of their salary from traffic tickets.
At the same time, there are not many cops making USD200K per annum in my municipality. And it's in flyover country, where everyone is clamoring for lower taxes. So I think it'd be a bit naive to think politicians wanting to score easy points with voters in cities, and even states, won't take the opportunity to extract a bit of revenue out of Big Tech.
Not saying it's right. Just saying that's how local politics work.
1 reply →
Bike Lanes have turned out to be an interesting edge case.
Waymos are currently dropping off and picking up passengers in a bike lane which is not legal (because it is dangerous) however many ride share drivers also do this. As somebody who is commonly a biker / pedestrian I am excited that AVs will likely make many things safer for that class of user. That being said, I do worry about how we encode these "social understandings" of laws. - A waymo I rode in on a highway was happy to go slightly above the speed limit - It seems at stop signs waymo prefers to be slightly aggressive to make it through rather than follow the letter of the law.
It seems silly that we have to teach robots to break certain laws sometimes but parking in bike lanes / yielding to pedestrians are laws that human drivers break all the time and I hope the mechanisms mentioned in the article prevent us from teaching robots to program anti-social but common behavior.
https://futurism.com/future-society/waymo-bike-lanes-traffic
As a cyclist and driver I figure you have to use some common sense. I probably break some regulations all the time like stopping where you are not supposed to briefly but being safe and not inconveniencing others is the main thing.
It's all pretty nuanced. I don't know where to draw a line.
For instance: Busy intersections with 4-way stop signs are an interesting example of how laws don't quite fit.
It's obviously important to get the order right since nobody wants to be in a car crash today. But the law (often -- we've got 50 states worth of driving laws and they aren't all the same) says something very specific and simplistic about the order: First-come, first-served; if order is unclear, yield to the right. Always wait for the intersection to be completely clear before proceeding.
That sounds nice and neat and it looks good on paper. It was surely at least a very easy system to describe and then write down.
But reality is very different: 4 way stops are an elaborate dance of drivers executing moves simultaneously and without conflict. For instance: Two opposite, straight-going cars can proceed concurrently works fine. All 4 directions can turn right, concurrently. Opposing left turns at the same time? Sure! While others are also turning right? Why not.
When there's room for a move and it creates no conflict, then that move works fine.
We all were taught how these intersections are supposed to work, but then reality ultimately shows us how they do work. And the dance works. It's efficient. Nobody gets ticketed for safely dancing that dance. (And broadly-speaking, a timid law-abiding driver who doesn't know the dance will be let through...eventually.)
The main problem with the dance is that it's difficult to adequately describe and write down and thus codify in law.
But maybe we should try, anyway.
The nuance for four-way stops is pretty simple. First come, first serve queue. Except you are allowed to jump out of order if you jumping out of order doesn't slow the people ahead of you down.
You’ve done a great job of explaining exactly how 4 way stops are terrible , and why they should be eliminated.
Only two countries make heavy use of them, so it seems less effort to get rid of them and the AI driverless world will be better without them
3 replies →
I read an article a while back that they made Waymo more aggressive, in the ways you mention, because they were quite annoying to other drivers when following the letter of the law. There is something to be said for following the flow of traffic.
I would imagine they would be able to revert back to more strict rule following once autonomous vehicles reach some level of critical mass and human drivers are needing to adapt to the AV traffic, rather than AVs needing to adapt to human traffic.
I wonder what happens legally if a biker plows into the Waymo, Casey Neistat style.
https://www.youtube.com/watch?v=bzE-IMaegzQ
I don't think you are legally allowed to deliberately crash into things even if they are annoying.
The law in that video - ticketing cyclists for not being in the bike lane - surprised me. There's nothing like that in the UK.
In SF it's legal for taxies to do pickups/drop-offs in bike lanes
I haven't seen any evidence Waymo does it anywhere illegal "just because rideshares do"
This is false. It is only legal in the rare event that a passenger requires curb-side access for accessibility/ADA reasons; any other use is still illegal. To quote SFMTA taxi training:
Only drop off in a separated bike lane if you have disabled or elderly customers who require direct access to the curb You may only pick up in a separated bike lane if the dispatcher tells you that the customer is disabled and must be picked up at a location that is next to a separated bike lane.
Taxi drivers often intentionally misstate this regulation because it’s more annoying to follow the law and find a legal place to stop so they pretend they are allowed to use bike lanes for any reason.
2 replies →
Yea, California civic code is pretty liberal with curbside parking parking, allowing it anywhere it isn't expressly prohibited, with signs declaring it so.
I live in Northern California, inland of San Francisco, and the city closest to me has a bunch of streets with bike lanes that are just painted onto the shoulder and otherwise are legally just a shoulder. Most of those streets also prohibit parking, but some don't, so parking is in the bike lane.
It gets really crazy in the denser parts of Southern California, where parking is sometimes not prohibited even when there isn't a shoulder, so parked cars full-on block a driving lane.
Good luck. This is going to be faceless machine v. faceless machine. I can't think of anything more useless in terms of fixing actual needs of society. In fact, I expect it to lead to shady business. Example: Waymos are all paid up at this intersection so traffic is allowed to safely ignore the rules for this month. Tesla Cybercab is still fighting us, so we continue to ticket them.
They haven’t been all this time? Damn — what a time to be a robot
I guess it's like patenents (when it's the same thing but comouter) or piracy (but it's model training at a FAANG), where tech just gets a free pass.
If the cars can incur liability, manslaughter would be next
What's gonna really be funny is the first time a state legislates that an AV company has to keep a bug in their software to maintain a municipal income flow.
_begins_? Like, before, they wouldn't get tickets?
Moving violations are issued to the driver, in person, so there hasn't really been a mechanism for issuing them to an operator.
Yes, and there's an example towards the end of the article.
As a Waymo (and other driverless car) supporter, this seems like an obviously good thing, right? I’m a little surprised this wasn’t possible before given the amount of regulatory scrutiny (correctly) applied to these companies.
Archive link in case of random paywalling like I got: https://archive.ph/xHMDO
> As a Waymo (and other driverless car) supporter, this seems like an obviously good thing, right? I’m a little surprised this wasn’t possible before given the amount of regulatory scrutiny (correctly) applied to these companies
Not necessarily. I went into a bit more detail in my own comment but it might be useful to think that when regulations are written keeping in mind multibillion dollar automobile companies, what the effect of those regulations on a person maintaining their own vehicles might be.
Consider that your Waymo got ticketed, but you had flashed it with a "no customer telemetry" firmware. Once Waymo gets the ticket, they flag your car as having "unauthorized" software and now the ball's in your court that the reason why your Waymo got ticketed has nothing to do with the telemetry feature that tells Waymo what radio stations you were listening to.
Also, when regulations are written keeping in mind multibillion dollar automobile companies, the ticket isn't going to cost $500.
I would hope any type of software modification would put more of the responsibility the owner.
1 reply →
Yes.
They should automate the fines by requiring the cars to send the video for processing through an enforcement program. Self-driving, self-ticketing. Maybe that would help force manufacturers to improve quality and not hide as much.
But I think the manufacturers should be hold responsible not the customers
Laws should be loser for autonomous vehicles with good safety records. No one is protected by preventing waymos from making rolling stops, and driving like a human Uber driver.
Ideally the fees would be similar to the Norway model, where some tickets are tied to the income of the driver, in this case the pre-tax earnings of the company that created the driverless car.
That can make sense (opinions differ) for individuals, but it's not like the company is advertising with "we get you there at 1.2x legal speed". They're not competing on that; they're not choosing to do this on purpose like an individual might choose to speed (for example because of economic incentives if their hourly price is high)
If they were, then it makes sense to fine them to some multiple of the benefit they got from this advertising tactic, but as it is, I don't see why it should be different from anyone else's ticket. The company isn't likely to enjoy a flood of this administrative work, besides the cost of the actual fines, so they'll work to minimise them anyway
They may not advertise “getting you there at 1.2x legal speed” but the sooner they drop you off, the sooner they can get another fare. Over a whole fleet it will add up to changing the size of the required fleet.
If getting a ticket one ride in a thousand is cheaper than deploying another 2000 cars to make up for the increased trip time I’d expect them to keep getting tickets.
I’m also not sure they don’t do it on purpose. Tesla self driving has an aggressive mode willing to speed and roll through stop signs. Those were deliberate, law breaking, choices.
1 reply →
I think a rich person would then do the rational thing and hire a cheap driver who also owns and operates their car.
Noone sane would be willing to assume basically unlimited liability for someone else's software.
Maybe that's good thing - some work for humans after the robots take over, albeit as human legal shields ;)
> they're not choosing to do this on purpose
So the AV software is so unreliable that it can't adhere to speed limits?
1 reply →
Assuming you divide it down to the earnings per car, that makes perfect sense. Of course right now they aren't making any profit at all, and by the time it is relevant it is likely that the cars will commit substantially no violations at all.
Driverless taxi services are all blowing through venture capitol, so does having a negative income mean that they get a payment, instead of a fine?
I think this could be a good compromise. Could have a floor value but the ceiling can vary accordingly.
Agreed. We need to look at reforming fines in general.
Fines should be scaled to income and the value of the vehicle and should exponentially increase for reoffense when in the same catagory of offense.
Isn't Norway only for drunk driving? Finland has it for massive speed excesses, but it is based on net taxable income taking out business expenses for taxi drivers, and Waymo is still negative.
If they become profitable you'd want to normalize by number of miles, unless you just want an incentive system to get more people on the road (extra drivers) and increase chance of humans suffering road injuries to boost employment in an internal service sector.
But even then coming out with a more efficient fleet than a competitor for higher margin would be penalized. You'd rather disincentivize skimping on safety for margin and not disincentivize better maintenance and fuel economy.
How do the police pull over a driverless vehicle? Who hands over the license and registration? How do you get it to sign the ticket?
Dunno what it's like in Cali but in the UK nearly all tickets are generated by automatic cameras and arrive in the post to the registered keeper of the vehicle.
You send robocop.
Why were we not doing this already?
This will be just another minor cost of doing business unless they are treated like human drivers in at least two other ways.
1) If theses companies get enough points on their license, their license is revoked. Not just for that vehicle, but for all of their vehicles. (The number of points would need to be adjusted for number of miles driven.)
2) Senior executives could be held criminally liable for vehicular manslaughter the way a normal drivers are. A death doesn't mean someone is going to prison, but their would be a police investigation. If an exec decided to ship a product with a known bug that lead to someone's death it should be treated with the same seriousness as a drunk driver killing someone.
A more pragmatic metric would be comparing deaths/mile for drivered cars against deaths/mile of driverless cars.
Why have they not been doing this already?
Why was this not already happening?
It’s about time!
I hope fees increase massively to incentivise change
This is weird
We should start charging them payroll taxes too per hour
I'm honestly surprised they weren't before
seems like real-word RL for AV. why not?
If there aren't serious consequences for driverless cars committing crimes (I mean jail time for executives serious), what's to stop someone for starting a hitman business?
We'll just run someone over with our "driverless" car and pay a fine - capitalism, baby!
That is great, they should also start ticketing human-driven cars that violate traffic laws too!
UPDATE (can't respond to the two subcomments below due to post throttling, so I'm updating this comment instead)
> the car is basically a taxi and the taxi service is to blame for any mistakes
@skybrian - Agreed! but if you read the article, the CA DMV is ticketing the manufacturer, not the operator.
None of my concerns hold if the operator was ticketed - infact, existing regulations are set up exactly that way, so no new regulation was even necessary. Something's not adding up.
> Right now, no one can independently own and operate an AV the way Waymo or Tesla does
@ourspacetabs - Sure but the regulation seems to be specifically addressed at the manufacturer, not the operator.
I would have no concern if the regulation was addressed to the operator. The article atleast doesn't imply that's the case.
---
> The state's Department of Motor Vehicles (DMV) has announced new regulations on autonomous vehicles (AVs), including a process for police to issue a "notice of AV noncompliance" directly to the car's manufacturer.
> Under the new rules, police can cite AV companies when their vehicles commit moving violations. The rules will also require the companies to respond to calls from police and other emergency officials within 30 seconds, and will issue penalties if their vehicles enter active emergency zones.
These are new frontiers in automotive regulation. Typically, if a car failed because of a manufacturer issue, the driver would be ticketed. For example: if Hyundai sold vehicles where the engine would explode around 50k miles and that caused an accident, the driver of the vehicle would be ticketed for it.
Now if we take the human out of it, it is Hyundai that would be ticketed for it. Insurance companies are certainly going to take notice and adjust their risk models accordingly.
I imagine there will be a lot of fingerpointing by the manufacturer towards customers.
In the worst case, this is the end of customers servicing their own autonomous vehicles.
If we imagine that most vehicles in the next 15 years will be autonomous, this would mean customers would have to handle regulation aimed at multibillion dollar companies, if they were to service their own autonomous vehicles, or give up on servicing their own autonomous vehicles entirely and just rent them instead.
Not sure I agree. The clear boundary here to me is who owns and is operating the vehicle. Waymo both owns and operates their vehicles, it’s a taxi service, you wouldn’t say a Waymo rider is operating a vehicle and therefore deserves the ticket. Right now, no one can independently own and operate an AV the way Waymo or Tesla does.
When that happens someday, then the ticket would go to the owner/operator of the vehicle - whoever bought the car. If you get a ticket due to something dumb your personally owned Waymo did, wouldn’t you pursue that case against Waymo separately, the same way you’d pursue Hyundai for selling you a car whose engine blew up after 50k miles?
It seems pretty reasonable to me that when you're not driving, the car is basically a taxi and the taxi service is to blame for any mistakes. The car manufacturer isn't just making cars anymore. It's providing a service.
Perhaps they could sell the car to a different taxi service, though?
maybe Tesla can put that weird robot that connects the charging connector to the car to use by building a robot that can give the police a hand to place the ticket into
Just now? Why the hell haven’t they all along?
I don’t disagree with needing some sort of consequence for bad driverless actions. But I distrust the motivation. Maybe California is just looking for more revenue sources after rampantly mismanaging their state and letting corruption and fraud continue.
[dead]