Tesla reportedly faces criminal probe into self-driving hype The US Department of Justice has reportedly launched a criminal investigation into Tesla and its boasts about its self-driving car technology. This probe could lead to federal criminal charges against the automaker, according to Reuters. The investigation kicked off, we're told, after a series of prangs in which Teslas in …
People who think "Autopilot" means autonomy, need to spend 5 seconds googling that topic.
And five seconds will just be long enough to skim a list in which the first four entries are not about autonomous movement of aircraft.
They're not even about Tesla, or Musk.
They're about some Microsoft product.
That's not really true. The first entry is some Micro$hit crap, but directly to the right of that is the Wikipedia entry, with the text
"An autopilot is a system used to control the path of an aircraft, marine craft or spacecraft without requiring constant manual control by a human operator. Autopilots do not replace human operators."
clearly visible on the google page.
The average person on the street would imagine Tesla's "autopilot" to be an autonomous car of some sort. They aren't going to understand the nuances or limitations of such a system.
Tesla were lying by omission by using the term and not educating people on the limitations. Even worse, they didn't even force the driver's attention to mitigate against consequences. So of *course* people started doing stopped paying attention to the roads to look at their phones, the pretty clouds and them BAM.
This is broadly correct, for the US. "Puffery" – advertising or marketing claims that would be understood as hyperbole or subjective ("the best-tasting floor polish money can buy!") by a reasonable consumer of ordinary knowledge – are not actionable under deceptive-advertising laws.
That said, there's nothing to stop the DoJ from pursuing charges; then it falls on Tesla to settle (no doubt with no admission of guilt) or try their luck in court.
And the DoJ could go after Musk personally, based on the public claims he made about Autopilot on Twitter and the like, where he was not acting in his capacity as an officer of the corporation.
...perhaps in the rest of the world, but in the good ol' U-S-of-A, NOT taking marketing hype seriously is tantamount to being considered un-American...or, rather, "un-ahMURRican", as approximately half the population would say.
"The best argument against democracy is a five-minute conversation with the average voter."--Winston Churchill
"Think of how stupid the average person is; now realize half of them are stupider than that.”--George Carlin
...perhaps in the rest of the world, but in the good ol' U-S-of-A, NOT taking marketing hype seriously is tantamount to being considered un-American...or, rather, "un-aMURRican", as approximately half the population would say.
What's up with people these days? I still love driving around in my manual shift mode of transport. I suppose having your hands free while the Tesla drives itself is that you can use your smartphone to take pictures as it crashes into other things. There's that.
My daily driver is a manual with no automation, however when I have the additional facilities on hire cars I'll use them when appropriate.
Cruise avoids the need to keep one eye on the speedo, adaptive cruise makes things even easier but requires specific knowledge and skills (i.e. if you're going to overtake the slower car in front pull out before your car starts to slow down to their speed).
Lane following I could take or leave to be honest. I've used it and it frees up the mind to concentrate on what's happening around you but you still need to be fully aware.
I think Ford/VW have done the right thing, L3 is as high as we're going to get for automation in general use on existing infrastructure.
"Cruise avoids the need to keep one eye on the speedo"
Not strictly necessary unless you intend to drive with the needle to the maximum line all the time (and that's not the safest way to drive). Thre's a host of other cues about speed -- the engine sound for the gear you're in, the rate at which roadside object pass and more. All that requires is attention to the act of driving, which is necessary to avoid accidents anyway.
Personally, I have a 20ish year old car with ye olde cruise control; which is to say that it maintains a steady speed at X MPH when set.
If i'm driving multiple hours then I have a habit of getting to a safe following distance (~100-150 yards) yards behind an HGV that's on cruise control at 60MPH and then set mine. I get 55+MPG cruising at that speed, and as people rarely want to pull in tightly behind an HGV then you'll be able to just quietly sit in the lane and cruise on.
Cruising at 70MPH meanwhile requires hundreds of lane and speed changes to the point it's hardly worth doing; the 10mph difference works out timewise as being 8 minutes per hour; it's also ~20% cheaper on fuel. Personally, i'd rather have the additional money than the 8 mins. Your view may differ.
I found that keeping to the speed limit as measured per GPS gives you a slightly higher speed than surrounding traffic, and on normal, not too busy roads that works for me (also because I don't HAVE to overtake, if I see something coming up behind me I'll nudge it down a bit to maintain safe driving distances). Driving with an ego is never a good thing.
That said, when it comes to traffic queues you'll find me mostly in the outer lane for the exact reason you just mentioned: much quieter, and it generally flows better because for HGV drivers coming to a full stop is costly so they tend to be better and flexing distances to keep rolling, and as a bonus I can see MUCH better what is going on ahead. As I hold all driver licenses (I got bored in my youth) I also understand that keeping with the flow helps the HGV drivers - you'll never see me, for instance, invade the space in front of a laden HGV which it needs for braking. Yes, maybe I'll get to my destination a few minutes later because of it, but at least I get there and relatively stress free at that.
I suspect you get a good MPG not just because of the lower speed (a trick most taxi drivers know as it gives less roll and wind resistance), but possibly also because you're in the vortex caused by the HGV in front of you. That's a theory, though, I don't know if someone has ever tested that but chaining of vehicles has long been the topic of many autonomous driving experiments. I think VW did something once where you had a string of cars automatically following each other closely, but the question becomes then who will accept taking the hit for others - you'd need some sort of change mechanism to go with it. In your case, the HGV has little choice :).
I suspect you get a good MPG not just because of the lower speed (a trick most taxi drivers know as it gives less roll and wind resistance), but possibly also because you're in the vortex caused by the HGV in front of you.
As far as i'm aware you'd have to tailgate (eg <20 yards) to get any effect on fuel efficiency from the HGV. At 100 to 150 yards behind an HGV i'm exceedingly sceptical that there would be any effect.
"As far as i'm aware you'd have to tailgate (eg <20 yards) to get any effect on fuel efficiency from the HGV. At 100 to 150 yards behind an HGV i'm exceedingly sceptical that there would be any effect."
The latter part is correct. The former, less so. It depends on speed, because it's noticeable that at the closest safe distance - 2 second rule plus a bit - behind a big SUV doing 90, my fuel economy is better than doing 90 on my own. When pulling out into the next lane over, you can feel the instant deceleration as wind resistance increases. If an HGV were doing 90, you'd probably notice a benefit a safe distance behind.
You can get a good idea of the wake size of different vehicles by looking at the tail of spray on rainy days.
FWIW Mythbusters did a segment on drafting behind semi trailers back in 2007. They got an MPG improvement of 39% at a following distance of 10 feet (3 meters) at 55mph.. However, tailgating a truck at 10 feet while driving i\at highway speed is neither legal (even in Boston) nor safe. The didn't measure savings at a safe distance (150 feet), but they did try 100 feet and got an 11 percent improvement in fuel economy.
https://www.autoblog.com/2007/10/28/mythbusters-drafting-10-feet-behind-a-big-rig-will-improve-mile/
the 10mph difference works out timewise as being 8 minutes per hour
A reasonable consideration for a one- or two-hour trip, provided you can safely go 10 MPH slower than most of the surrounding traffic. In the US that's often risky – even the semis are typically cruising at or above the posted limit if they can.
For a 15-hour drive, where the difference means 17 hours instead, it's less appealing.
I completely disagree. Take the M25, for example: the camera-enforced limit is 70, the so-slow-it's-dangerous speed is 60, so you aim to sit halfway between the two but even slight variation in speed takes you into either ticket or coffin territory.
It also takes significantly more concentration to listen to engine revs etc. I was astonished how much more relaxing motorway driving became when I finally got a car new enough to have cruise control. I just sit there with my thumb on the speed controller and keep an eye on the gap to the vehicle ahead.
I've done long (15-18 hour) drives without cruise control, and I've done them with cruise control; and let me assure you, cruise control is much easier on the right ankle.
And, frankly, maintaining a consistent speed at highway speeds, on a low-traffic highway, in a good-quality car, is not particularly easy. In my Volvo there's very little engine noise to begin with, and other external sounds are substantially diminished, so distinguishing between, say, 65 MPH and 75 MPH by ambient noise is not really feasible. Maybe you can gauge your speed precisely by parallax with roadside objects, but I can't – not to the degree that would keep me safe from speed traps, certainly.
For what it's worth, I hate adaptive cruise control. In my opinion, cruise control is there to maintain the speed I set, not whatever speed the nitwit who pulled into my lane four car lengths ahead thinks I should be going.
"I still love driving around in my manual shift mode of transport."
Well, that's really what the problem is. I like driving for fun. Not a big fan of driving to get somewhere, because that's usually boring. Especially for motorway trips and so-on I'd much rather read a book while the car does it for me. And for commuting, I'm looking forward to the day my self driving 'car' is actually a self driving mobile bathroom, so my morning shower etc can happen while the car negotiates traffic - it would give me an extra half hour in bed, I reckon.
I think you mean "a problem that you don't personally have".
The world is full of people who cannot drive, for any number of reasons - the young, the old, the blind, the drunk, and so on. Autonomous cars will bring huge benefits to all of those, without the high costs of the current solution, which involves having someone else dedicated to doing the driving.
Then there's the question of utilisation. Most cars spend 95% or more of their time parked up, doing nothing. Hugely wasteful, and hugely disruptive in large towns and cities. Especially true of second cars in families. The utilisation of autonomous cars should be way higher than this, allowing the world to contain many fewer cars. This is a good thing.
GJC
"The world is full of people who cannot drive, for any number of reasons - the young, the old, the blind, the drunk"
The young can learn to drive. As for the old, blind and drunk, see point about funding public transport or subsidising taxis.
"Most cars spend 95% or more of their time parked up, doing nothing."
See point about public transport and home working.
To be fair, nothing says autonomous vehicles couldn't be rented, and indeed we already have such on offer in some places.
So it's really a contest between "vehicle driven by a person that you own or rent" and "vehicle driven by a machine that you own or rent". I personally find the latter concept tiresome, but I admit arguments can be made for it. I'm not yet convinced they're compelling.
Given my car experiences when the kids were growing up I would not want to inflict what they do to a car on anyone else :).
I see that a lot with rentals (whose prices also have gone up dramatically), sometimes you get one that hasn't had the cleaning it needs. Normally I'll hand that straight back but it depends on how much time I have.
If and when we have autonomous 'taxis' it will be interesting to see how they are priced. I suspect it will be at least demand based, I do wonder if a vehicle on the way to you might be reassigned if a better/more profitable trip comes up. You might then get an alert with the offer to accept a higher charge to still get the same pickup time or take a new later slot.
Definitely seen this idea in fiction at least.
And then, at the other end of the market, you could choose to use a multi-occupancy taxi that would not necessarily take the most direct route, in exchange for a lower cost. Effectively a small free-roaming bus, but one that picked you up and dropped you off exactly where you wanted to be, rather than 15 minutes walk away.
There's lots of possibilities. Some might happen, some won't. We'll see.
GJC
Elon's bloviating about autopilot were 100% fraudluent and completely undoable with the technology he had and even more undoable with the technology he has now since he's stripped every sensor but cameras from Tesla's cars. Tesla's 'autopilot' and 'self driving' have been sold for years by Musk as hands off, let the car do it - the term 'autopilot' entirely implies that. But his simple DNN can never deliver it.
So yes, it was entirely fraudulent and he should be prosecuted for it.
Last week, Musk said Tesla would soon release an upgraded version of “full self-driving” software, allowing customers to travel “to your work, your friend’s house, to the grocery store without you touching the wheel”
As far back as 2016 Musk described it as “probably better” than a human driver.
A video currently on the company’s website says: “The person in the driver’s seat is only there for legal reasons. He is not doing anything. The car is driving itself.”
Damned by his own mouth - nothing new there for him.
See also:
https://www.theguardian.com/technology/2022/oct/26/tesla-criminal-investigation-self-driving-claims-sources
Despite the money spent, they still cannot solve the fundamental problem of how a A.I vehicle solves real life issues such as whether to avoid a person by crashing the car and other moral dilemmas (a.k.a. the trolley bus problem). This is because what is touted as A.I isn't. It is excellent set of pattern recognition algorithms tied to simple decision tree. The problem is it can only deal with issues thought up by the designers and do not handle the variability of real the real world. This is where wetware still is superior to such systems
One solution suggested is that the A.I. does the easy stuff such as managing the vehicle on things like motorways and then give control back to the driver in situations outside its parameters. Unfortunately we have already seen issues where this can go horrible wrong. An air France flight from Brazil was on autopilot (much easier on an aircraft since you don't meet many people at 20,000 ft). The air speed sensors had frozen up causing an issue the autopilot was not designed to handle so it requested the pilot take control. The pilot did, misread the situation, panicked, tried to pull the nose up of a already climbing aircraft, stalled it and crashed it. The vehicle equivalent would be you happily cruising along the motorway in A.I mode, half asleep, because you have nothing to do. The A.I then telling you of some issue, you waking up, assessing the situation in a second and making a calm valid response. Multiply that situation a million times and we have potential disaster on our hands
However hopefully some of the technology will go to improve driver augmentation. These systems have an advantage that they are always on and alert, but they need to work with the driver not be in control.
As for Tesla, it is another example where the hype has come up with the realities and while I am all for moonshot projects, there has to be honesty as to what is achievable
One solution suggested is that the A.I. does the easy stuff such as managing the vehicle on things like motorways and then give control back to the driver
In theory that sounds fine but in practice there's an unsurmountable problem - delay - by the time a driver has noticed the car want's them to take over and then analysed the situation a couple of seconds will have passed and the problem will probably have resolved itself, probably by crashing.
"It is excellent set of pattern recognition algorithms tied to simple decision tree"
Or not so excellent pattern recognition algorithms? There have been numerous reports (including accident reports) of the "AI" misinterpreting things ranging from low bed lorries as bridges to off ramp intersection markings as lanes, and also failures to recognise (consequently crashing into) stationary obstructions ranging from police cars to (in one case) a large black and yellow chevroned crash barrier end marker.
The so far (and probably permanently) insoluble problem is that the "AI" can't understand anything (that is: extrapolate what it detects to its potential implications more than one stage away). As that's what the human brain does as a matter of course, there's a huge difference between the kinds of "intelligence".
One solution suggested is that the A.I. does the easy stuff such as managing the vehicle on things like motorways and then give control back to the driver in situations outside its parameters.
Lane keeping and adaptive cruise control already create attention issues with drivers, AI driving will only make that worse because when the AI is going to ask for attention it means you're already in a sticky situation. To build up situational awareness from scratch takes minutes you don't have, and if you're maintaining situaltional awareness constantly you don't really need an AI.
I'm all for driver assist system where they make sense (but I am absolutely against mandating that these things are by default on, whoever made that law should be sacked IMHO), but whoever has been selling this as an argument to stop paying attention deserves all the legal problems that can be thrown at them.
This post has been deleted by its author
Well, Tesla do publish some data (although there doesn't seem to be any for 2022 - not sure if they have stopped) but the last batch from 2021Q4 is:
In the 4th quarter, we recorded one crash for every 4.31 million miles driven in which drivers were using Autopilot technology (Autosteer and active safety features). For drivers who were not using Autopilot technology (no Autosteer and active safety features), we recorded one crash for every 1.59 million miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 484,000 miles.
from: https://www.tesla.com/VehicleSafetyReport
Tesla numbers are a mathematical rounding error in comparison.
Of the about 300 million cars in America only about 1 million are Teslas. So Teslas are less than 0.4% of cars in America.
Of course they will contribute a tiny value to the total, the question is is the total a) disproportionate and b) involve "autopilot".
Not as far as I'm aware, no. I guess maybe the total mileage on autopilot comes from Tesla, but afaik it comes from submissions to regulators under penalty of perjury, so they would be unlikely to lie - massage the figures a little, sure, but not outright lie. We can be fairly confident that as a ballpark figure it's right, and ballpark figures are all we're dealing with when it comes to global averages too.
It isn't an implausible result, so I don't think there's any need for a great deal of skepticism here. It isn't very hard to do better than the average untrained driver set loose on the roads, which is why we don't use that as the standard that must be reached.
The thing about the current generation of semi-self-driving cars that isn't usually mentioned is that they are already significantly safer than the average human driver, based on a global average. They are safer than the average US driver too. Not as safe as the average driver in the handful of countries around the world that have proper driving tests.
Really all this says is that the standard of human driving is so absurdly low that it's not hard to do better.
To me, there are only two methods of automated driving. Say lane change and anti-tailgating assistance, and FULL AUTOMONOUS DRIVING. By the time an unattended driver is alerted something is wrong to the reaction time for the driver to take command a crash happens.
The consumer is obviously confused by all this, and the auto companies (and without a doubt some politicians) are taking advantage of all of this..
The Tesla cameras have a range of 250m apparently, which is only 8 seconds at 70mph, however I don't think they start allowing for changes that far away unless they are static obstruction. Even then, as you say, how long will it take the car to decide it cannot cope with something that it has seen and let the driver know? Even driving manually normal drivers can take two or three seconds to notice and react to changes on a motorway, if they are just letting the car drive it could be longer still. Then they will need to work out what the car is reporting and finally react to it. As you say, partial autonomous driving is not a great idea.
"I don't think they start allowing for changes that far away unless they are static obstruction"
Actually, they seem exceptionally bad at recognising static obstructions. There are numerous accident reports of teslas driving straight into them, and at least one of a tesla actively accelerating into one.
There also seems to be a number of cases of Tesla driving in to the back of motorcycles.
Here at CNN and I think there's been another more recently.
As much as I dislike Tesla's hype machine, it seems very unlikely there is anything that can be successfully prosecuted. It is generally OK to make all kinds of outlandish claims in advertising.
https://en.wikipedia.org/wiki/Carlill_v_Carbolic_Smoke_Ball_Co
https://en.wikipedia.org/wiki/Puffery#Federal_Trade_Commission_definition
That said, I hope it goes to court. The defence would be fun - pretty sure they'd argue no-one in their right mind should take any of Musk's claim's seriously, and possibly also that 'full self driving' refers to Tesla owners fully having to drive themselves...
That doesn't really work given the rate of crashes even when the system is abused. Ignoring the need to charge the car, he'd likely starve or die of thirst, and then the corpse would liquify, well before a crash happened.
It's a win for humanity either way, but I'm not sure it would demonstrate a lot about the cars.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
