Driving impaired in any way shape or form is dangerous.

https://www.dailycamera.com/2016/06/09/affidavit-teen-driver...

As is driving unimpaired.

https://www.stripes.com/us-army-soldier-three-others-serious...

Yes, automobiles are very dangerous and it'd be great if we used them a lot less.

That said, driving impaired is even more dangerous and I'm totally fine with tools to crack down on it. Smoke or drink all you want, but do it at home or with someone to give you a ride.

Oh I agree with you, I just think linking to an example of a teenager who was late getting home and got into an accident when he had been smoking marijuana earlier as if it were supporting evidence makes our position look cheesy.

nobody should ever be allowed to drive ever then

Perhaps - automobile-related deaths outnumber gun-related deaths about 30:1.

And that's not to mention all the deaths from diseases associated with air pollution.

Agreed, robot cars when?

Cars using ML will at times have to answer to philosophical questions such as trolley problem.

Honestly, no they don't. If the car is following all applicable laws, no one should ever be in any danger. If someone has violated the law, they have inherently accepted the risks of their decision, and thus the car should protect the occupants.

Someone tied to the street, doesn't apply because (in my state) the car should be moving at a speed that it would be able to stop in time. Someone thrown from a bridge into a cars way isn't one either, the car should try to protect the occupants again, because that's the case where it'll do the least amount of damage to other nearby cars. In the latter case, Humans wouldn't be able to make a moral decision in that amount of time either, and thus, should default to doing the least amount of damage.

> If someone has violated the law, they have inherently accepted the risks of their decision, and thus the car should protect the occupants.

Well, then you are going to need a different ML ruleset here in The Netherlands because bicyclists are protected by law. In the case of an accident with a car, the car driver is (almost always) at fault. That being said, since july 2019 it is illegal to operate a smartphone from your bicycle which, at times, was at odds with the earlier mentioned law.

I assume other countries in the world are going to have different (perhaps in eyes of some peculiar) exceptions to what is deemed normal.

The whole problem with your argument is this it is all "nice, in theory".

> should default to doing the least amount of damage.

"Should", indeed. The question is whether all of these ethical related aspects have been thought out. As the above suggests not only from the US (or whatever state you are in) PoV but with taking into account the entire world and the entire world's ethics and law. Heck, you might have to switch rulesets when crossing a border. Either way that is a very, very complex ruleset you're going to have to work with. And I doubt it is being done...

Does smoking marijuana impair driving skills though? The assumption is yes... but is that actually the case?

Maybe the kid in the story you linked crashed due to his speeding to reach home by curfew and not because he smoked marijuana that night.

It would make sense that it does, at some level of consumption, just as with anything else that messes with your brain, including alcohol, prescription drugs, lack of sleep and so on.

What appears intuitively true is not always so. Certainly, anything at "some level of consumption" would be an impairment, but I'm not sure it's established that marijuana impairs driving ability under typical usage scenarios.

Given how dangerous automobiles are even for non-impaired people, I'd rather err on the side of caution, especially given that consuming is not something people need to do.

I prefer to err on the side of data before it is mandated that people be tested for marijuana at traffic stops in the name of "caution".

Makes sense to follow similar protocols to alcohol:

https://www.oregonlive.com/politics/2015/07/marijuana_and_dr...

In other words, this isn't something that anyone who gets stopped goes through. You get tested if you're driving erratically.

Driving is a privilege. It's easy enough to avoid doing it while high.

> Makes sense to follow similar protocols to alcohol

Only insomuch as those techniques have nothing to do with alcohol specifically and can be applied to any situation, including sleepy drivers.

> In other words, this isn't something that anyone who gets stopped goes through. You get tested if you're driving erratically.

Not true. You could be pulled over for a variety of reasons, anything from speeding to a broken tail-light or just because the cop thinks you look like a stoner and is betting he'll be able to bust you. From the article you linked:

> The officer evaluates the driver. Some signs are obvious: bloodshot eyes, candy bar wrappers and what Iwai calls the "typical Beavis and Butthead laugh." Others are more subtle -- for example, slightly protruding veins on the person's neck.

So the officer makes some subjective judgements that will allow him to administer the test and whether it comes back positive or negative, we have not established that the driver was actually impaired and this is critical because the test would then allow the officer to imprison the citizen even if they were not driving erratically. The suggestion that you'll only be pulled over for driving erratically is not true.

Who's to say a 20 year old driving high is any more or less dangerous than an 80 year old geriatric? Drunk driving is definitively provable as dangerous, high driving is not.

you cannot conclude from that article that driving under the influence of cannabis is dangerous.

The literature on this is .... not great.

..yes? Is it some trick question?

No, all the published articles I've read come from testing people who've been in a MVC, who were also at fault. With plenty of confounding variables as well.

I'm specifically looking for a dose dependent study that shows level of THC intoxication has positive correlation with driving impairment. With a large sample size. Hopefully from some group not looking to prove something.

I just know how a high works on me, and I'd never drive in that condition. It would be really bizarre if a drug that effects behavior/motorics would make no difference on driving. A cursory search shows the issue has been studied for a while too:

https://www.nap.edu/read/18942/chapter/9#114

Certainly there are people who would say it has no effect on them. Met people who thought a few beers makes them better drivers, too.

The National Highway Traffic Board's studies on the matter are remarkably inconclusive.

no shortage of toxicology literature proving it. check ramaekers' publications. he sticks high folks in real cars on closed tracks to see how they do.

Gonna need a source for that, very interested. DDG gave me nothing

Since there’s no shortage I’m sure you can provide some solid links to statistics on marijuana-impaired driving deaths?