The vehicle was travelling 20mph faster than it's own "safe clear ahead" zone allowed for. Simply ignoring it's own safety limits and travelling at a higher rate of speed is a design flaw.
Designing a large public road autonomous vehicle test and not taking into account an easily predicted human vulnerability that's been known in other automation experiments for decades is a project design flaw.
Whether or not the driver is at some fault, I think the court will be capable of determining that. Arizona's willingness to completely disregard the mountains of negligence on part of a giant corporation is disturbing.
If the car is going over it's speed limit, the driver should react.
I see that some people think Uber should have added systems to control the safety driver. I think that's a fair point, but it doesn't take away the drivers responsibility. They knew the car isn't perfect. It's their job to take over control whenever that happens.
> I can't see how this is not clear cut negligence.
1) The driver is a woman.
2) While I think the driver does bear some fault here, they don't bear all the fault. Uber designed an unsafe system that relied on an unnatural amount of vigilance from a single person while simultaneously discouraging that vigilance [1]. They didn't design the car to shut down when one of it's critical safety components (the driver) was not operating correctly, and they didn't even give that driver amphetamines or something to increase their vigilance to artificial levels.
[1] Basically: pay close attention to a boring process while doing absolutely nothing for hours on end. I'm pretty sure that's a classic "humans suck at this" task.
This isn't a truck driver falling asleep from exhaustion caused by aggressive scheduling. You don't "accidentally" take out your phone when your job is looking ahead. That's deliberate negligence.
If you want to require eye sensors to detect distraction, by all means, pass a law about it. Maybe include regular non-AI cars too. #1 cause of accidents over here.
Statements like this seem to indicate a failure to understand what the difficult part of the task was. In short: a boring monitoring task that was practically unnecessary 99.99000% of the time, but absolutely critical maybe 0.00001% of the time. The typical-case lack of necessity would strongly re-enforce engaging with distractions over time (i.e. distraction more rewarding, and nearly all of the time no negative feedback of any kind).
> This isn't a truck driver falling asleep from exhaustion caused by aggressive scheduling. You don't "accidentally" take out your phone when your job is looking ahead. That's deliberate negligence.
I'm not saying there was no negligence here on the part of the driver, just that Uber itself was at least as negligent if not more so for designing the system in the way it did. Focusing too much on the driver is an error.
Honestly, watching a show was egregious, but I wouldn't be surprised if the drivers would end up falling asleep on a regular basis if all distractions were removed from their environment. Half paying attention to something I wasn't interested in has always been the best way to get me to fall asleep.
Doing that for five minutes is easy.
Can you do that for eight hours? Day after day? Without a single lapse in attention?
This is a much harder job then the truck driver has - because he constantly has to make microadjustments, to correct for road conditions.
People's brains don't work the way you think they do.
I am saying that passively watching the road for hours on end is much harder than actively driving.
What makes you confident that you are actually good at it, and are not the victim of Dunning-Kruger? Do you regularly find yourself in the process of stopping your self-serving car from crashing into things?
Or has your car simply not crashed yet?
All the 3 of them? Among millions of Tesla vehicles out there and the hundreds of millions of miles driven? Is that even a considerable risk when compared to the general (non-zero!) risks of driving?
For me personally I simply know when I watch the road and when I don't. For some people this might be hard, for others - not so much. I am aware of when I pull out my phone or distract myself and when not.
But are you just watching or are you seeing the road? How many times have you taken control from your car doing something stupid and dangerous?
Unless the answer to that second question is 'I do it all[1] the time, and I'm batting 20/20', what makes you confident that you'll catch the next instance?
[1] If that's really the case, you should probably short TSLA, it doesn't sound like their car can safely operate.
Yes, I think I would be able to do this job safely. Not everyone has problems with focus to such an extent that you can't help but be on the phone when you need to be paying attention.
> and they didn't even give that driver amphetamines
Yo, what? IIRC, most *amphetamines are either illegal in the US or like FDA Class 1 substances or something like that, and you just want to throw them at people operating > 1 ton machinery? That seems like it could have an even higher risk for danger.
That's true, and my comment was mostly sarcastic (but a little serious). Amphetamines are given to military pilots to increase alertness [1]. They're also used to treat attention deficit disorder (stimulants also increase focus in healthy people, see coffee). Some years ago I read that they're also popular in some parts of Asia not for recreation, but to help workers focus on boring, repetitive tasks (and they apparently can even make that kind of work seem "fun"), but I can't find the exact article again.
The kind of job these "car monitors" are expected to do is so unnatural that it's ludicrous that someone could be expected to succeed at it perfectly, unassisted.
[1] https://aviation.stackexchange.com/questions/8885/do-militar...
Also, no one said the driver has to do nothing. They can safely listen to audiobooks or talk on the phone.
I get that it’s boring but fail to see how that’s an excuse for not paying attention (or watching streaming video) to a dynamic system you’re expected to actively engage with.
Many train systems are largely autonomous and far more static and yet we still expect train operators to not kill someone on the tracks.
How do you feel about lifeguards as a job? Failure of pools and beaches to design a good system?
Train conductors do not require second-level reaction time. A train cannot stop in five seconds, no matter how quick the conductor is. People who are stopped on train tracks get killed, and the train operator is very rarely to blame.
It's the basic job description. If she couldn't do it, she shouldn't have taken the job.
It's unreasonable to hold someone accountable for a "self-driving car" that suddenly decides at a split-seconds notice that it can't cope driving.
Of course this is extra bad because it's an experimental car, but it's the same in my opinion with those Teslas on the road now that do the same thing.
It is not unreasonable at all. She had one job to do - look at the road. She failed it because she felt that her entertainment was more important than doing her job. She picked up her phone and started streaming videos. She failed at her one job, plain and simple. She knew everything about the job and still chose to watch some videos and risk lives.
The experimental car shouldn't have been on the road at all if the only thing separating it from killing people is someone who is expected to maintain concentration for hours/days while simultaneously not actually doing anything.
Then, if we start with the expectation that a human in the loop is necessary for live testing, we could have two. It reduces the impact of independent failures. It also adds some social pressure to avoid negligence. It also provides overlapping coverage, as a lifeguard in this thread has pointed out.
Humans are fallible. Accept that, and design systems to be safe despite individual errors.
Driving a car is not the same aa monitoring a self driving car. There will be differences in attentiveness. Stop equating them.
In my opinion, safe FSD cars are not possible with current technology on existing public roads - so either the technology has to improve by orders of magnitude, or the roads need to be modified significantly.
The gung-ho experimentation that is going on in public is in my opinion very dangerous and should be stopped, and this case is a perfect example of why.
Both the road infrastructure and other drivers are too crappy and unpredictable for FSD cars to be viable and safe.
Also this business of holding someone who is not actually controlling the vehicle accountable is ridiculous, and is surely something which will be proven in court eventually.
If you go on a long multi-day road trip, you might even end up relying on a passenger's feedback in a moment of tiredness.
>A backseat driver may be uncomfortable with the skills of the driver, feel out of control since they are not driving the vehicle
Are you saying that the meatbag in a "self-driving car" is equivalent to a backseat driver?
Because if you are, then "self-driving cars" will never be viable because the humans inside don't actually trust them.
The claim I'm disputing is the one that says people are somehow incapable of paying attention to the road for extended periods of time unless they are in control of the vehicle.
I might agree that a reaction might be jerky and panicky, but then again, they would by definition be so regardless, due to the unexpected nature of accidents
That sounds like a nice safe system.
For the record, I don't agree. I feel that the average person who would choose to use a self-driving will tend to be overly relaxed and trusting, not what I'd call a backseat driver.
The kind of person who is a backseat driver wouldn't trust the car's driving and would prefer to control it themselves, and thus wouldn't be using the self-driving feature in the first place.
Humans are very bad at paying attention for a rare event and having nothing else to do 99.999% of the rest of the time. This is surely well known to the people who set up this system in the self driving car.
Don't take a boring job if you can't handle it. The required basic level of continued attention is not a superhuman skill. It's not a job for everyone but it's not exceptional at all.
Driving a truck is in no way comparable to what this lady was tasked to do. A truck driver is performing an active task (operating a truck), this lady was supposed to be performing a passive task (monitoring a truck's driving). Active and passive tasks are very different beasts.
But you're on to something with your comparison to pilots. Planes are mostly automated. When that automation fails, the pilots often lack the situational awareness to avoid a crash when they have to take over (for an example, see https://99percentinvisible.org/episode/children-of-the-magen...).
How many times do you think this driver was distracted with the phone when doing their job? I bet it wasn't just this one time that they happened to crash.
Pilots crash after hours of boredom because debugging problems in the air is hard, not because they were negligent leading up to the problem (vast majority of the time, anyway). And still by and large they succeed, it just goes unreported in the news because business-as-usual. This driver on the other hand got into an unrecoverable situation because of their own behaviour.
Dude was paid to pay attention. He didn't. It wasn't because he got fatigued, it was because he was deliberately doing something else, by choice
The guy screwed up. Uber knew he almost certainly would. I won't say the guy isn't responsible at all, but pretending this is purely on him and Uber bears 0% of the responsibility is just silly. Humans are not capable of anything and everything simply because they are provided pay.
If she had been negligent, i.e., too tired to notice the pedestrian, that would have been a different matter. If she had been paying attention most of the time but inadvertently reached over to drink some coffee, that would have been a different matter.
But she wasn't too tired, and she wasn't momentarily reaching for a drink; she was simply choosing not to pay attention to her job so she could do something else instead.
I'm saying that the inattention was basically a given. Yeah, at some point the driver went and grabbed a phone. It may have been fiddling with junk in the car, or their fingers, or day dreaming to the extent they were unaware of the world around them.
Yeah, they grabbed a phone. Is that really any different from any of the above? I don't think so. Particularly considering how ingrained the reflex is to grab and fiddle with a phone when bored is. I wouldn't be surprised if they never consciously decided to pick the thing up.
The task was simply unreasonable. Uber should have foreseen this problem and implemented systems to aide and monitor their drivers.
Aside from this, if the car had been driving over the speed suitable for the conditions at the time the fall person had plenty of time to note that and take over, in that case it was more like huh, the paint is drying but it looks like it's going to rain .... long time looking like it's going to rain ... oh no should have put the covers on!
But I'm betting the need to take over and drive is not properly given to the fall person either.
It is not at all fair to make the comparison you are making.
I can't see how this is not clear cut negligence.