Tesla's 2M car Autopilot recall is now under federal scrutiny

The IIHS has repeatedly shown that Tesla airbags are the best in class. Deploying airbags when not needed can do significant harm to occupants. Not to mention total vehicles for no good reason.

There isn't a need to wonder when independent data gatherers have already answered the question.

I have no knowledge or opinion about what thresholds various auto makers have relative to each other, beyond the fact that they probably all meet regulations.

But I do want to point out that ~nobody would buy a car that deploys airbags for "pretty minor bumps". Airbag deployment is a serious event, it can injur or in rare cases kill a person itself.

That's how airbags work in general.

A slight bump won't set off your airbag in any car.

Tesla frequently loses telemetry in serious and fatal crashes [1]. For a specific example you can see here [2] listed as 13781-3074 where the fatal crash was detected by a complaint rather than telemetry. Of reported conclusively fatal accidents, only ~45% were detected by telemetry.

From that alone, we can conclude that their telemetry is bad at fatal crash detection. However, when we contrast it against their regular crash reporting where ~90% are telemetry, we can further conclude that either crashes are massively under-reported or fatal accidents are uniquely under-reported (even ignoring the unknown base rate vs reported rate problem). In either case, the massive unexplained discrepancy in the most visible, fatal crashes demonstrates that their data collection procedures are grossly inadequate to conclude any positive safety statistic.

The fact that Tesla repeatedly markets safety statistics based on grossly inadequate data collection procedures without any attempts to identify the true, higher rate, and without qualification is criminally repugnant.

[1] https://static.nhtsa.gov/odi/ffdd/sgo-2021-01/SGO-2021-01_In...

[2] https://www.washingtonpost.com/technology/interactive/2024/t...

> I've often wondered if Tesla is gaming the figures by only deploying the airbags in really bad crashes, whereas other manufacturers might deploy the airbags for pretty minor bumps.

So even as a Tesla "hater", but as a paramedic, first generation airbags were very simplistic. Essentially, "if speed > x mph and collision = true then deploy".

Nowadays, advanced airbag systems do calculations based on speed, G forces, angle of intrusion, lateral movement, etc., and decide "will the airbag system cause or risk more injury?", i.e. "can the passenger restraint system (i.e. seatbelts) do sufficient work to minimize injury".

So you can get relatively low speed collision with heavy lateral movement causing a deployment, but some other deployments at 30mph not deploying airbags.

All that being said, I do not believe Tesla (or any other manufacturer) so describe accidents only in terms of airbag deployment. Other factors like speed, G force felt, did seatbelt pretensioners activate, should all be sufficient to categorize as an accident.

Autopilots in planes maintain altitude and speed and no landing or evasion.

But full self driving was a broken promise.

I can see you can potentially argue (unsuccessfully IMHO) that it is marketing fraud in inducing purchases of the vehicle, but I don't buy it that a single person who actually owns the vehicle and have driven on Autopilot for more than a week is confused by the abilities of the system because the name is Autopilot. People text and drive _all the time_ on their ICE carriages without Autopilot too. Thus the claim that renaming it would have saved a life, let alone lives, is simply sensationalism.

It was the name _and_ constant absured marketing claims, not just the name

To be fair, I don’t think Autopilot is well defined in any governing literature. We’re in the wild west.

The underlying issue isn't the name, it's the mismatch between marketing claims, actual performance, and what it lets you "get away" with.

They released a system that will engage on almost any road and provides a firm steering force that will steer without driver interaction.

Unlike some competing systems, they did not include dedicated hardware for monitoring driver attention, or restrict operation to separated highways.

Ever since, they have been trying to layer on increasingly annoying nags to jiggle the wheel, or use the low res interior camera to try to determine driver attention, to try to work backwards from the fact that they didn't provide for proper driver monitoring to begin with or build a system that discourages zoning out.

Not sure. Anyone who used it for a mere hour would understand its limitations immediately.

Otherwise - all systems can be abused.

To wit, "incident was so catastrophic that restraint systems could not deploy"/"incident was so catastrophic it could not report" is not theoretical. In fact it is the case in the majority (yes, majority) of reported conclusively fatal crashes [1].

[1] https://static.nhtsa.gov/odi/ffdd/sgo-2021-01/SGO-2021-01_In...

Don't forget all the drunk, high & exhausted drivers too.

100 people are killed every single day on the roads in the US. Day in and day out.

I generally agree it does great on freeways, but not in urban situations where there are tight corners and lots of parked cars, or rural areas with twisty roads and farm equipment. I'd turn it off in construction areas too.

For sure but also consider those autopilot apis are written by impatient coders, distracted proj managers, and with crappy specs.

Not all "miles driven" are the same. Autopilot is going to be engaged on highway driving which already has a lower accident rate per mile driven. So that biases a simple comparison.

Edit: and even within highway miles, autopilot will be enabled more in the "easy" stretches and turned off in roadworks etc.

On similar roads, in similar traffic, in similar weather.

If you look at fatalities or define “accident” as cases where somebody had to visit a doctor or something like that, you also need to compare to similar cars. Most self-driving cars are relatively new and relatively large, and both decrease those (probably even accident rate because they statistically have newer brakes and tires and have more active safety features)

It’s far from trivial to decide whether some program is a better driver than the average human driver if the program gets the newer bigger car and can decide to cop out in cases where we do not expect humans to do so.

Probably not, no.

I believe 2014 is the model year of that Tesla, it seems like that particular crash was in 2018. https://incidentdatabase.ai/cite/320/

It says

> A 2014 Tesla Model S

the vehicle is a 2014.

The photo is of a Jan 2018 crash

https://www.ocregister.com/2019/09/04/ntsb-tesla-autopilot-l...

> Just hammer the people who abuse them with reckless driving charges.

That surely helps the now dead people they crash into. The manufacturer can prevent deaths by adding safety features, so it should to a reasonable extent.

A lot of Tesla marketing over the years has been super deceptive to the point that a lot of people assume the car is capable of fully self driving itself. I've had a model 3 since 2018 and the number of people that have asked "do you ever drive yourself?" is too damn high.

So Tesla actively does everything they can to make people thing the system is way more capable than it really is. Then shows a small popup the first time you use it that tells you "j/k it doesn't really do any of this, goodluck". I feel like eventually they'll be held accountable for it, but the bureaucracy moves veryyyy slowly.

> But it's also not Tesla's responsibility to nanny people

It literally is.

> Self-driving systems shouldn't need all these "are you paying attention" bells and whistles.

It's not self driving at all. That's the problem.

If you sold alcohol called "Safe Drive Juice", implied in advertisement that you can drive safe while drinking it, even made you a better driver, in tweets and press releases said that this is the future and everyone should use it, and if they did they'd be safer drivers, it doesn't matter if you put a note at the bottom that said "doesn't actually let you drive safe, please don't drink and drive"

I remember back in 2013 when I first drove a Tesla, they were really pushing the self driving features at my local dealership. I can't help but feel like they were overselling their capabilities, and their staff could well have made a gullible person believe the car really does drive itself perfectly.

> it's also not Tesla's responsibility to nanny people

Yes, it is. An L2 driver assist system must ensure that the driver is always actively supervising, and Tesla's systems failed to do that. That's why they recalled 2 million cars.