The numbers they gave is more than a doubling, while they were quite clear about the limitations of their data. That includes not knowing how many AVs are on the road, which is something that the companies should be able to provide.
Also look at the claims that AVs have not caused death or serious injury. (I noticed that they did not claim that AVs have not caused injury.) That is great, except I would not expect any given corporate vehicle fleet, autonomous or human controlled, to have a record of causing death or serious injury over a short period of time. The sample size is just too small.
Either way, there's not enough data to prove anything. On the other hand, we have a group with a clear conflict of interest (the makers of AVs) up against a group which has noted concerning incidents but has not been provided with the data they need even when it should be available. Then we have a third group who are being asked to let an experiment on the general public proceed. I doubt that it would pass many academic ethics committees, but if you have money, well, go ahead!
In February they hit 1,000,000 miles cumulative for all previous years. Last week they hit 3,000,000 miles.
Cars kill about as many Americans as the Vietnam War, but every year. Aside from guns, I can't think of anything else so deadly that we give to our children as they enter adulthood.
It is common, it is not part of the news cycle. We care more about the unlikely terrorist. People are bad with numbers.
Personally, I think it would a huge ethics violation to not be running tests of autonomous cars (not talking Telsa toy driving stuff).
A cumulative 3,000,000 miles is nothing. The US Bureau of Transportation Statistics estimates there are over 3,000,000,000,000 miles of vehicle traffic per year.[1] The estimated number of fatalities is around 50,000. There is no way to assess whether there will be fewer or more traffic fatalities with current AV technology given the limited amount of data.
As for the unlikely terrorist bit, just in case you weren't around when 911 happened: even mathematically inclined people were shocked. Not only was it the most lethal attack on American soil (nearly 3,000 dead), it was a foreign attack. People genuinely didn't know what was going to happen and were living in fear for a while. Unfortunately, some people still carry those fears to this day. Even though the numbers don't back them up, I wouldn't be so quick to
dismiss their emotions and I certainly wouldn't attack them for the all too human mistake of misattributing risk.
Humans average 60,000,000 miles between fatalitys [1]. It is a average of 1.25 million miles between injurys. 3 million miles is not a statistically significant amount of data to make a valid estimate, not even close.
That is not to say that safe testing can not proceed or that they are being unsafe in their testing or validation process, but people are bad with numbers, so it is important to realize the actual magnitude of the status quo and what we are comparing against.
Anecdotally, I feel much safer around AVs as a pedestrian or cyclist than human driven vehicles. Way more predictable, especially post pandemic when it seems like half of drivers have some behavioral issue.
Driving condition matter since they vary from place to place. It matters for people, and it most likely matters for autonomous vehicles (either due to training data sets or direct programming of traffic regulations). To choose a mundane example, that is admittedly more likely to affect people, consider how many people try to make a left turn into a (North American) roundabout or who park in a bike lane. (Sometimes it is deliberate, but sometimes it is an out of town person who has never dealt with it before.)
That said, I would expect a San Fransisco decision to be based upon San Francisco data.
Anyone seen the rushing dash cam videos where pedestrians are throwing themselves at cars in order to get legal liability payouts. Not that I'm accusing all of the reports here of being something like that, but I wonder if robotaxis are going to result in something like that and actually make the entire driving world a lot less ethical.
You will always have opportunists. This article is not about that. The specific incident described by the article is about a car going through caution tape and getting caught up in a fallen overhead powerline. There's very little chance of payouts there, though there probably should be if considerable damage was caused. When it comes to interference with emergency responders, I also doubt it being a case of people looking for payouts.
A lot of AV enthusiasts seem to be painting the world as against them, when the reality is that people don't like being unwilling experimental subjects. If there was proof that AVs were safe, I could imagine people jumping on them in droves. Why would one reject having a vehicle where you have the option to drive yourself or have the driving being done for you? (I realize this article isn't about that scenario.) On the other hand, the tech industry's mantra of moving fast and breaking things - something that existed in practice long before the likes of Facebook - has bred an incredible amount of distrust. That distrust has only grown as it has shifted from the technology itself into grand social experiments.
> If there was proof that AVs were safe, I could imagine people jumping on them in droves
These services have waiting lists.
San Francisco is simply a city that can’t help but distract from its endemic problems by torching those who bet on it. (That said, I blame these firms for choosing San Francisco.)
Also look at the claims that AVs have not caused death or serious injury. (I noticed that they did not claim that AVs have not caused injury.) That is great, except I would not expect any given corporate vehicle fleet, autonomous or human controlled, to have a record of causing death or serious injury over a short period of time. The sample size is just too small.
Either way, there's not enough data to prove anything. On the other hand, we have a group with a clear conflict of interest (the makers of AVs) up against a group which has noted concerning incidents but has not been provided with the data they need even when it should be available. Then we have a third group who are being asked to let an experiment on the general public proceed. I doubt that it would pass many academic ethics committees, but if you have money, well, go ahead!