Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Blog Tesla Crashes into Cop Car After Launch of NHTSA Autopilot Investigation

This site may earn commission on affiliate links.


A Tesla operating on Autopilot hit a Florida Highway Patrol car Saturday, according to a report.

The Orlando Sun Sentinal reported that a trooper was helping a disabled vehicle in the westbound lanes of I-4 near downtown Orlando. With his emergency lights on, the trooper was helping the driver when the Tesla hit the left side of his car. There were no injuries.






The National Highway Traffic Safety Administration (NHTSA) announced early this month an investigation into Tesla’s Autopilot feature.

The agency pointed to 11 crashes since January 2018 where Tesla models operating on Autopilot “have encountered first responder scenes and subsequently struck one or more vehicles involved with those scenes.” The agency said the accidents caused 17 injuries and one death.

“Most incidents took place after dark and the crash scenes encountered included scene control measures such as first responder vehicle lights, flares, an illuminated arrow board, and road cones,” the investigation summary said. “The involved subject vehicles were all confirmed to have been engaged in either Autopilot or Traffic Aware Cruise Control during the approach to the crashes.”

The investigation includes about 765,000 Tesla vehicles in the U.S., applying to the entire lineup since 2014.

Image: Flordia Highway Patrol

 
Last edited by a moderator:
Interesting. Current wheel torque sensing disables AP after 3 events. So you're being more permissive with eye sensing.
Remember, this 3 seconds would apply to looking at the screens in the Tesla too. So picking music, entering navigation, turning on seat heaters... Those would all trigger this.
Tesla already has a dashcam and records before crashes.
This would not magically decline accident numbers. It might make "crashes on AP" decline. But it might make overall crashes increase, since you're now having distracted drivers operating with no AP assistance, instead of some. We're not causing the car to stop here, just not use AP.

It's fascinating that we'd all rather people driving around with no AP assistance after we know they are distracted. Shows just how much we trust AP to actually avoid an accident if we don't think it can go 3 seconds without solid monitoring. I personally dream of a day, long before AP is L3+, where the actual usefulness of AP is that it can reasonably assist someone while they do something else for 3 seconds.

It's telling about where Tesla believes AP development is if they are spending a lot of energy implementing and training driver monitoring instead of improving the reliability of AP. The thing they flat out don't need when they are L3+. It tells you how far away they believe it to be.

Looks like Tesla agrees, and is going in the right direction regarding the cabin camera. The numbers I posted were very loose, but I'd like to see the system function similar to Blue Cruise in terms of monitoring. A sensible threshold can be chosen that's not too restrictive, but ensures more driver engagement.

Unconscious drivers is another area where Autopilot is a little behind. Yes, the car would eventually stop as long as there is no torque to the wheel, but if the driver is slumped over on it, it may not until it reaches an obstacle. Here, the cabin cam could literally be a life saver. Also, the dashcam won't relay enough info about the state of the driver before a crash.

As far as torque sensing, what's employed at the moment is too liberal and doesn't do enough to prevent the driver from getting dangerously distracted. Used in tandem with the cabin cam would allow for a more effective system, IMO.

And yes, people will moan about privacy, but safety trumps privacy. If someone runs into a vehicle causing serious injury or death because they 'thought' autopilot would always do what it's expected to', the family won't want to hear anything about 'privacy'.
 
Last edited:
Unconscious drivers is another area where Autopilot is a little behind. Yes, the car would eventually stop as long as there is no torque to the wheel, but if the driver is slumped over on it, it may not until it reaches an obstacle. Here, the cabin cam could literally be a life saver.
There have been hundreds of Tesla accidents with fatalities. Are you aware of a single one where the driver was slumped over the wheel? Seriously, if you care about actual statistical reduction in fatalities, this is not the place to be going.

Plus, if medical conditions are what you are worried about- we should be encouraging MORE AP use, so that AP has a chance to control the car to a stop, not less AP use.

Is the goal here less accidents overall, or less accidents while AP is engaged? Are you sure booting distracted drivers out of using AP will serve that goal?

but safety trumps privacy
Tell that to the airline pilots union who have successfully fought off cockpit cameras for decades despite them being on the NTSB's "most wanted" list of safety enhancing technologies.

Safety trumps nothing. Safety is a tradeoff like everything else. The safest car is one that never moves. If safety trumped privacy, medical, driving, and other records would be public.
 
  • Like
Reactions: Hank42
There have been hundreds of Tesla accidents with fatalities. Are you aware of a single one where the driver was slumped over the wheel? Seriously, if you care about actual statistical reduction in fatalities, this is not the place to be going.
These are not medical emergencies, but they drive home the point. These people are asleep and Autopilot is still engaged. This isn't safe. I'd rather Autopilot disengage and bring the car slowly to a stop rather than keep going and possibly collide with something or someone. And if Tesla could go one step further and have the car pull the left or right (depending on which it's closest to)off the roadway as much as possible, even better. Even if one life is saved, it's worth it.



Plus, if medical conditions are what you are worried about- we should be encouraging MORE AP use, so that AP has a chance to control the car to a stop, not less AP use.

Is the goal here less accidents overall, or less accidents while AP is engaged? Are you sure booting distracted drivers out of using AP will serve that goal?
Would you trust a distracted driver with AP engaged over an alert driver with AP disengaged? I wont. This is not a critique of Autopilot nor is it about discouraging its use. Rather, it's about improving driver engagement. A distracted driver is a potential accident waiting to happen. Autopilot might reduce the odds of one happening, but treating an L2 system like a level 3 or 4 is not wise.

Tell that to the airline pilots union who have successfully fought off cockpit cameras for decades despite them being on the NTSB's "most wanted" list of safety enhancing technologies.
Voice cockpit recorders are just as invasive, so I don't see the point of fighting off cameras. What does a pilot have to hide in the cockpit? Is he / she going to take their clothes off and go at it with co-pilot? We can travel safely today because of the contributions of voice and data recordings.
If safety trumped privacy, medical, driving, and other records would be public.
The DMV can request medical records about a condition if the situation calls for it. And they have access to your driving record. They don't need to be public.
 
Last edited:
Gonna need a source for that. Green's videos show much more than cell use as something the system is trained on.


I also wonder how it can differentiate looking at your phone vs the center screen, given it can't really see your hands in many positions.


Yet ths car company lets you use a web browser on the screen while driving...
I was going off what I recalled Green tweeting about when he described his experience with it. Where he said something in regards to phone usage triggering warnings more than other stuff. Seems pretty straight forwards as phone usage is definitely something Tesla will want to curb while using AP. Of course I could be misremembering.

I don't have any experience with the driver monitoring as I've never seen it show up on any of the release notes with the my Model 3 with HW3, and with Radar.
 
These people are asleep and Autopilot is still engaged. This isn't safe. I'd rather Autopilot disengage and bring the car slowly to a stop rather than keep going and possibly collide with something or someone. And if Tesla could go one step further and have the car pull the left or right (depending on which it's closest to)off the roadway as much as possible, even better. Even if one life is saved, it's worth it.

Tesla already detects when the driver is asleep via torque monitoring, and comes to a stop. Explain how these people defeated that torque sensing... And then ask why they can't defeat driver monitoring. Green has already shown a picture hung near the camera will beat it.

Voice cockpit recorders are just as invasive, so I don't see the point of fighting off cameras. What does a pilot have to hide in the cockpit? Is he / she going to take their clothes off and go at it with co-pilot? We can travel safely today because of the contributions of voice and data recordings.
Good point. Call up ALPA and make this point to them and I'm sure they'll reverse their last 20 years of successfully fighting cameras. I mean, if you don't see the point, there must not be one.

Like I said, NTSB agrees that it would improve safety. My point is ALPA has successfully argued that privacy is more important and prevented it from becoming law.

If safety trumped privacy, medical records would be public because then everyone could know if their neighbor was vaccinated or had a different contagious disease.
 
Tesla already detects when the driver is asleep via torque monitoring, and comes to a stop. Explain how these people defeated that torque sensing...
Good question. Thanks for making my argument for me. They are clearly asleep and the cars are still driving. Having a second monitoring system could help in this situation.
And then ask why they can't defeat driver monitoring. Green has already shown a picture hung near the camera will beat it.
No one said anything about not being able to defeat a system. Plus, in your example above the camera would record the picture. This country has no shortage of Darwin Award nominees.

A medical condition or someone falling asleep for whatever reason is what we're talking about here. As in an unplanned situation. Apples and oranges.
Good point. Call up ALPA and make this point to them and I'm sure they'll reverse their last 20 years of successfully fighting cameras. I mean, if you don't see the point, there must not be one.

Like I said, NTSB agrees that it would improve safety. My point is ALPA has successfully argued that privacy is more important and prevented it from becoming law.
Legal bribery (aka lobbying) often has its tenticles somewhere in the mix. It wouldn't be the first or last time doing the right thing wasn't popular.
If safety trumped privacy, medical records would be public because then everyone could know if their neighbor was vaccinated or had a different contagious disease.
It is easy enough to assume the other person isn't vaccinated and take the necessary precautions. I don't need to see their medical records. As far as a contagious disease is concerned, seeing their medical records wouldn't necessarily make a difference if they hadn't seen a doctor yet. And if they had seen one, then most likely they were treated already or are in treatment.
 
Last edited:
Next step will be law enforcement threatening punishment if people are caught doing stupid stuff to circumvent safety measures.

Maybe a combination of eye monitoring and wheel torque will be the ticket, they'll continue adding safeguards until defeating them is too onerous to be worthwhile. And then anyone found attempting to defeat them will face fines and hits against their license.
 
Next step will be law enforcement threatening punishment if people are caught doing stupid stuff to circumvent safety measures.

Maybe a combination of eye monitoring and wheel torque will be the ticket, they'll continue adding safeguards until defeating them is too onerous to be worthwhile. And then anyone found attempting to defeat them will face fines and hits against their license.
Explain to me why Autopilot is a safety system, and not a Advanced driver-assistance system (ADAS).

Intentionally disabling safety systems is already illegal.
 
Explain to me why Autopilot is a safety system, and not a Advanced driver-assistance system (ADAS).

Intentionally disabling safety systems is already illegal.
ADAS systems are safety systems. Backup cameras, ABS, LDW, AEB, even driver monitoring are all ADAS. What is the point of a system that "assists" a driver but doesn't assist in safety? They all need to be designed knowing their unexpected operation has a safety impact at the bare minimum: Advanced driver-assistance systems - Wikipedia

Advanced driver-assistance systems (ADAS) are groups of electronic technologies that assist drivers in driving and parking functions. Through a safe human-machine interface, ADAS increase car and road safety. ADAS use automated technology, such as sensors and cameras, to detect nearby obstacles or driver errors, and respond accordingly.

As most road accidents occur due to human error,[2] ADAS are developed to automate, adapt, and enhance vehicle technology for safety and better driving. ADAS are proven to reduce road fatalities by minimizing human error.[3] Safety features are designed to avoid accidents and collisions by offering technologies that alert the driver to problems, implementing safeguards, and taking control of the vehicle if necessary. Adaptive features may automate lighting, provide adaptive cruise control, assist in avoiding collisions, incorporate satellite navigation and traffic warnings, alert drivers to possible obstacles, assist in lane departure and lane centering, provide navigational assistance through smartphones, and provide other features.[3]

Many of these technologies are legally mandated (ABS, Backup Cam, Traction Control, Stability Control, etc) because of their proven safety benefits. However, I'd like you to show me a law that is common to all states saying that driving a car with one of these broken or intentionally disabled is illegal.

If AP is not a safety system, tell me why Tesla publishes a safety report that argues AP is 10X safer than a human, and at what feature set it will be a safety system?

Customers can be forgiven for considering it a safety system given this is how Tesla describes AP:

Autopilot​

Autopilot advanced safety and convenience features are designed to assist you with the most burdensome parts of driving. Autopilot introduces new features and improves existing functionality to make your Tesla safer and more capable over time.
All Tesla vehicles produced in our factory, including Model 3, have the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver.
 
Last edited:
ADAS systems are safety systems. Backup cameras, ABS, LDW, AEB, even driver monitoring are all ADAS. What is the point of a system that "assists" a driver but doesn't assist in safety? They all need to be designed knowing their unexpected operation has a safety impact at the bare minimum: Advanced driver-assistance systems - Wikipedia



Many of these technologies are legally mandated (ABS, Backup Cam, Traction Control, Stability Control, etc) because of their proven safety benefits. However, I'd like you to show me a law that is common to all states saying that driving a car with one of these broken or intentionally disabled is illegal.

If AP is not a safety system, tell me why Tesla publishes a safety report that argues AP is 10X safer than a human, and at what feature set it will be a safety system?
You are correct! ABS, Backup Camera and others started as CONVENIENCE features. They were then PROVEN to be safer and then moved under SAFETY features.
Autopilot is on it's track and this answers the question at the end of your post.
 
Last edited:
Even if we just look at the driver engagement side of things, stuff like steering wheel torque and eyeball monitoring are safety systems that would be required during use of the ADAS and circumventing those would be an issue.

There are surely people who research this stuff, some of them are likely employed or funded by the NHTSA. They'll do studies and determine the best way to leverage the safety benefits of this technology while limiting the downside risks.
 
You are correct! ABS, Backup Camera and others started as CONVENIENCE features.
Please, explain to me how ABS, stability control, or traction control are convenience features even when first introduced? Are you saying the first advertisements for ABS said "the convenience of not needing to modulate the brakes on slippery surfaces"?

This is like saying the airbag is a convenience that just makes a crash less uncomfortable. Just because they were proactively implemented by manufacturers before they were legally required or the safety benefits were fully proven doesn't make them "convenience" features until they are legally mandated.

ADAS systems are safety systems, even at first introduction. Their failure can have major safety impacts, and their whole rationale for existence is reducing human harm via technology, even if they are not the sole solution on the vehicle (mirrors vs backup cameras). This is why when Tesla introduced AP2 in 2016, the phrase "safety level substantially greater than that of a human driver" was in the literal first paragraph on the page and the word "safe" existed on the page 7 times, and the word "convenience" was never used. AP did not start as a convenience feature.

They were then PROVEN to be safer and then moved under SAFETY features.
Tesla's own safety report says AP is 10X as safe as the average driver, and Elon calls current AP "superhuman" on the highway. Has it not been proven?
 
Last edited:
Please, explain to me how ABS, stability control, or traction control are convenience features even when first introduced? Are you saying the first advertisements for ABS said "the convenience of not needing to modulate the brakes on slippery surfaces"?

This is like saying the airbag is a convenience that just makes a crash less uncomfortable. Just because they were proactively implemented by manufacturers before they were legally required or the safety benefits were fully proven doesn't make them "convenience" features until they are legally mandated.

ADAS systems are safety systems, even at first introduction. Their failure can have major safety impacts, and their whole rationale for existence is reducing human harm via technology, even if they are not the sole solution on the vehicle (mirrors vs backup cameras). This is why when Tesla introduced AP2 in 2016, the phrase "safety level substantially greater than that of a human driver" was in the literal first paragraph on the page and the word "safe" existed on the page 7 times, and the word "convenience" was never used. AP did not start as a convenience feature.


Tesla's own safety report says AP is 10X as safe as the average driver, and Elon calls current AP "superhuman" on the highway. Has it not been proven?
As you wish, replace convenience with "differential". For market penetration, competition. Stuff was good and became standard. Do you not see the same path happening right now with aspiring self driving systems?

ADAS are not safety systems. They CAN BE safety systems.

My understanding is that all safety systems are ADAS, but not all ADAS are safety systems. ACC does not ADD safety to the drive, but it has to be SAFE ENOUGH to be used.

Has AP been proven safe to use?
I think we should agree that the answer to that question is "yes, while supervised"

Safety does not mean autonomy. ODD must be respected.
 
ACC does not ADD safety to the drive,
Maintaining a gap to the car in front of you instead of just driving into it doesn't add to the safety of a drive? Like you say, ADAS and automation are very different, but just because it still requires an attentive driver at the wheel doesn't mean it can't be a safety feature. In fact, one of the primary reasons for "assistance" technologies is ones that acknowledge drivers sometimes do become distracted and assist with that (like TAAC, LDW, AEB, etc all do).

ADAS are not safety systems. They CAN BE safety systems.
Basically every description of ADAS on the internet disagrees with you.

Advanced driver-assistance systems, are technological features that are designed to increase the safety of driving a vehicle. LogisFleet explains that when properly designed, these systems, referred to also as ADAS, use a human-machine interface to improve the driver's ability to react to dangers on the road.
Advanced driver-assistance systems (ADAS) are groups of electronic technologies that assist drivers in driving and parking functions. Through a safe human-machine interface, ADAS increase car and road safety.

NHTSA's page on "Driver Assistance Technologies" says this (and it includes ACC):
Driver assistance technologies hold the potential to reduce traffic crashes and save thousands of lives each year. In 2019, 36,096 people died in motor vehicle crashes — many of these crashes were tied to human error. Learn more about driver assistance technologies, how they can help you, and what you should know about these technologies when buying your next vehicle.

Show me a common industry definition of ADAS that says something can be "convenient" with no proposed impact on safety. What is "Advanced Driver Assistance" if not something that increases safety? What is "convenience" except the idea of reducing workload in the interest of increasing a driver's ability to react?

Also, Elon has been talking about AP as a safety technology forever. He said this in 2016 about AP1:
No, he responded. "There won't ever be zero fatalities," Musk said. "There won't ever be zero injuries. It's about improving the probability of safety, which is the only thing that's ever possible."

And finally, Tesla's own AP support page:

Autopilot is an advanced driver assistance system that enhances safety and convenience behind the wheel. When used properly, Autopilot reduces your overall workload as a driver. Each new Tesla vehicle is equipped with 8 external cameras, 12 ultrasonic sensors and a powerful onboard computer provide an additional layer of safety to guide you on your journey.
Autopilot is a suite of driver assistance features that comes standard with the purchase of a new car or can be purchased after delivery, and brings new functionality to your Tesla that makes driving safer and less stressful. Available packages include:
How does Autopilot work?
Built on a deep neural network, Autopilot* uses cameras and ultrasonic sensors to see and sense the environment around the car. This robust sensor and camera suite provides drivers with an awareness of their surroundings that a driver alone would not otherwise have. A powerful onboard computer processes these inputs in a matter of milliseconds to help make your driving safer and less stressful.
Even Tesla calls blind spot monitoring a safety feature:
Is there blind spot detection?
To improve safety and increase confidence when changing lanes, cars built since October 2016 will display a red lane line when your turn signal is engaged and a car or obstacle is detected in your target lane.

Seriously, do you think a reasonable person would read all of this and think "Oh yeah, AP is just about convenience?" Tesla always lists safety before mentioning convenience/stress/workload.
 
Maintaining a gap to the car in front of you instead of just driving into it doesn't add to the safety of a drive? Like you say, ADAS and automation are very different, but just because it still requires an attentive driver at the wheel doesn't mean it can't be a safety feature. In fact, one of the primary reasons for "assistance" technologies is ones that acknowledge drivers sometimes do become distracted and assist with that (like TAAC, LDW, AEB, etc all do).


Basically every description of ADAS on the internet disagrees with you.




NHTSA's page on "Driver Assistance Technologies" says this (and it includes ACC):


Show me a common industry definition of ADAS that says something can be "convenient" with no proposed impact on safety. What is "Advanced Driver Assistance" if not something that increases safety? What is "convenience" except the idea of reducing workload in the interest of increasing a driver's ability to react?

Also, Elon has been talking about AP as a safety technology forever. He said this in 2016 about AP1:


And finally, Tesla's own AP support page:




Even Tesla calls blind spot monitoring a safety feature:


Seriously, do you think a reasonable person would read all of this and think "Oh yeah, AP is just about convenience?" Tesla always lists safety before mentioning convenience/stress/workload.
All of what you said is true. We go to Driving school and learn how to follow rules and drive safely. Therefore anything that assists with driving has to be just as safe. Until they are, which is Tesla's case, they need disclaimers.

ACC > CC. I agree, but that's not the point. My point is that ACC should not be trusted and therefore is less reliable than an attentive meat driver. Therefore it's a safe feature, not a safety feature.

You don't have to convince me that ABS and the nanny things are positive safety additions. They overall are. And (exceptions apply?) disabling them is illegal. See NHTSA Standard 208. By logic, since it's legal (and encouraged often times) to disable Tesla's Autopilot, it shouldn't be considered a Safety feature. Just like ACC. But using them properly would indeed increase safety.

I encourage you to look as Tesla's videos from within your car. I'm not being a smart ass, I think they're really well done and even though I didn't learn anything from them, they provide a nice B2C vision of how Tesla interprets/markets the features they add to the cars. This is not verbatim, but they say this about autopilot: "After taking a trip with autopilot, you won't ever want to go on trips using another car again. Sorry, but it's true."
 
And (exceptions apply?) disabling them is illegal. See NHTSA Standard 208.
FVMSS 208 relates to the sale of a new car into the stream of commerce.

This says nothing about an owner modifying their car after they own it. These would be state laws. Show me a state law that says "operating a vehicle which was originally equipped with ABS with ABS non-functional or disabled is illegal." Heck, New Hampshire still doesn't require you to use a seatbelt! Many cars allow you to disable stability control/traction control with a button even though they are legally required for new cars. Why would they allow you to do something illegal?

ACC > CC. I agree, but that's not the point. My point is that ACC should not be trusted and therefore is less reliable than an attentive meat driver. Therefore it's a safe feature, not a safety feature.
Ok, then what is a "safety feature"? I mean, an airbag needs an attentive driver, it shouldn't be relied upon to save you and it's less reliable than not crashing. Neither should the crash structure of the vehicle. Are brakes the only safety feature of a non-L4+ autonomous car? Why does Tesla list AEB, SCW, FCW, and Auto High Beams as "standard safety features"? They don't make the car autonomous, and the manual warns you not to rely on them.

I encourage you to look as Tesla's videos from within your car. I'm not being a smart ass, I think they're really well done and even though I didn't learn anything from them, they provide a nice B2C vision of how Tesla interprets/markets the features they add to the cars.
And I'd encourage you to look at Tesla's web pages and manuals to see how they market Autopilot and safety features, and just how often they list safety in conjunction with Autopilot. Beyond that, I don't see how a video which is only available in a car you already own to be considered "marketing". How do I find these videos as a prospective buyer? They're not on the AP support page or the AP marketing page. A video on traffic light control is though, which shows the use of autosteer on city streets, a direct violation of Tesla's own manuals telling you Autosteer is not to be used on city streets. Which document/video am I supposed to believe?

The idea that Tesla has any kind of clear marketing/messaging on Autopilot can easily be disproven by just looking at a different bit of Tesla marketing/messaging.
 
Last edited:
FVMSS 208 relates to the sale of a new car into the stream of commerce.

This says nothing about an owner modifying their car after they own it. These would be state laws. Show me a state law that says "operating a vehicle which was originally equipped with ABS with ABS non-functional or disabled is illegal." Heck, New Hampshire still doesn't require you to use a seatbelt! Many cars allow you to disable stability control/traction control with a button even though they are legally required for new cars. Why would they allow you to do something illegal?


Ok, then what is a "safety feature"? I mean, an airbag needs an attentive driver, it shouldn't be relied upon to save you and it's less reliable than not crashing. Neither should the crash structure of the vehicle. Are brakes the only safety feature of a non-L4+ autonomous car?
What are we discussing? Damage reduction and control = increased probability of survival and lower physical harm. I'm agreeing with you on this.

But I think our discussion here is whether AP should have stopped for the emergency vehicles or not.
I'm saying no, but it would be good if it did. The system is not ready yet and my understanding is that it will.

I'm not really up to discuss state laws because in this country, that is wild. But generally speaking, if you drive without a seatbelt or with one of those clips to silence the buckle, it's illegal, so the exception proves the rule.
 
But generally speaking, if you drive without a seatbelt or with one of those clips to silence the buckle, it's illegal,
The clips are not illegal. They disable a chime, and there is no state law requiring the chime to drive on the roads.
Driving without a seatbelt is illegal in 49 states.
But notice that is a SPECIFIC thing. A seatbelt.
There is no generic law that says "disabling a safety device is illegal." Only disabling a specific safety device.
To do so generically would require.... Defining safety devices.
It would also mean that that if any safety device broke, you would be required to pull over to the side of the road and have the car towed to be fixed immediately. For people that want safe use of AP to be about personal responsibility and think NHTSA looking into AP crashes is a government overreach, wanting the government to make any disabling of any "safety" device illegal is mind numbing logic.

What we are discussing here is if AP is a "safety device." Of course it is. It's ADAS, and ADAS technologies are safety devices. You can't play with words here and try and say that when AP fails, it's no big deal because it's a convenience device, not a safety device, and the human was supposed to be paying attention, so the failure of the technology is no big deal and the driver was an idiot.

An airbag, ABS, or stability control would all fall into the same category. You're saying that if a crash occurs, the airbag does not go off, it's no big deal as long as the manual says "don't rely on the airbag"? The driver was just a dummy for getting into an accident and over-relying on the airbag?

And of course, if you want to argue that AP is just convenience then the weight devices on the steering wheel are not illegal, because they are just tricking a convenience device, not a safety device. The same will be true of tricking the driver monitoring related to AP. Why would it be illegal to override a convenience feature? Are you saying driver monitoring that is only active when AP is engaged is a safety device, but AP itself is not?

Tesla is going to get absolutely nowhere with Autopilot if they aren't on the path of acknowledging it as having some level of reliability, and Tesla having some kind of liability when it does not perform well. How does anyone think they will ever get to L3+ if the ongoing attitude is "it's convenience!" AP is squarely in the same category as all non-autonomy devices right now- it is supposed to help with the task of driving, and be a backup to the failures that real humans have. What is AP's value if it requires an equally attentive driver as without, and it offers no backup/harm reduction to human failures? Just a neat technical demo? If so, when do you think Tesla will move on from that and start officially acknowledging it as a harm reduction technology, and that the conversation can change from "that driver was an idiot for not obeying the clear manual" to "What a shame that AP wasn't able to save that real human driver that made a mistake that time, but on average, it's still a net positive and hopefully will continue get better?"
 
Last edited:
  • Like
Reactions: DanCar