Self-driving cars are surveillance cameras on wheels
schneier.com
schneier.com
(This is separate from this footage eventually being used to sell ads, as well as it being used non-consensually to "investigate" other crimes. I think Schneier is right to highlight concerns around the latter.)
We’re already down a dangerous path that isn’t really slowing down. Privacy in the digital age needs to be akin to no unlawful search and seizure.
But instead the ones at the top have us bickering against each other about more mundane issues, in the grand scheme of things
It's interesting they could have the most up to date mapping of anyone, I wonder if they plan to do anything with it.
No worries, Amazon has an excellent reputation, they would not think of doing shady things for money.
That doesn’t have to mean “government cameras”; I’m open to any solution that achieves those goals, including stricter licensing, urban redesign, harsher fines, etc.
There are many of those, in many formats, with many different behaviors. I'm not sure what lack of surveillance the OP is talking about.
But there does exist a high reluctance of transit rule enforcers to enforce transit rules, together with a high bias into looking only at the rules aimed at protecting the people inside the car, mostly at places where the car is creates the least risk to people outside of it.
Concentrated power with cameras-- one party controlling a lot of them, and doing automated aggregation of information-- has been bad for privacy.
Lots of cameras in public without the abuses of aggregating information seems best.
It is nice mind experiment that cars could theoretically spy on everyone, but hitting some practical problems, like who is paying for traffic from a car to some datacenter which will be in high megabytes every second? Who pays for petabytes of data to store every hour? Who pays for massive server clusters to search through those data when you need them?
For development prototypes, sure. For real production pieces it is unlikely to store that much data if any, because storage is expense and expense is lower margin and margin is thin.
It's pretty safe to assume that the basic combo of sensor outputs+vehicle logs will be retained (subject to retention period) for the foreseeable future, both of which have potential privacy issues. All the relevant companies have lots of internal discussion about this, but very little makes it into the public sphere.
PMs are given a target price and a target cost. They will do everything in their power to get under it because their job depends on it. Those drives are going to be the bare minimum necessary for storing what they need to (not what they want to) and will likely have a very short lifespan overall.
So given that context, how likely do you think it is that they're logging it all to terabytes of onboard flash?
But also, lifetime analysis is standard practice during the design phase.
My Toyota has an optional subscription to allow its data connection to be used for roadside assistance services, but the connection is always there regardless (and is used by Toyota to push updates to the car's touchscreen software without me paying for any subscription).
Alternatively, a hacker can kidnap you. :D
Or make your car stop on the freeway [1].
[1] https://www.wired.com/2015/07/hackers-remotely-kill-jeep-hig...
The lock on my front door only has to be strong enough to stop whoever wants to physically approach my house and attempt to pick it. The lock on my home network needs to be strong enough to stop anyone in the world from attempting to hack it.
> There's always going to be a way for malicious people to do harm.
This is an example of statement, that while technically true, is so useless it might as well not be. It would be like saying that humans have been killing each other since before nuclear weapons existed, so there is no reason to worry about multi megaton thermonuclear warheads. They are different levels of risk by orders of magnitude.
When you’re connected to the internet the number of adversaries willing to try goes up by probably 6 orders of magnitude. Not only that, many of them will be living in a regime where there is no possible way for them to get punished by doing it to you.
Ransomware brought significant amounts of income to the North Korean government.
These are not extreme hypotheticals. They’ve shut down fucking hospitals to shake down people for money. Once the opportunity is great enough, they’ll have no qualms stopping people’s cars from starting to ask for a little crypto currency.
These are totally different scenarios.
Like I get bulletproof cars, but the stopping power of a single electric car at 4000kg accelerating to 120mph in a matter of seconds changes the calculus completely and that’s before accounting for battery explosions or chemical fires.
Doing that with a single car today without remote execution is a suicide attack, but an army of hackers? Good lord, this sounds like a doomsday scenario.
Even if for some reason you had to have close proximity to the car to disable these settings doing at a place that is a disadvantage to the driver helps your position.
In the US, I am astounded at what is commonly accepted now. Not all that long ago, most people would have thought even the level of surveillance that we accept right now was an impossibly orwellian thing that would never have been accepted.
Yet here we are. I even see people arguing regularly here on HN that these are good things, when just 30 or so years ago, these would have been things that we would have rioted to prevent.
I have come to think that there is no amount of oppression that Americans can't be convinced into.
I consider it a fourth amendment issue. The government has no business remoting into your car, rummaging though its data, and seizing control of your property.
Maybe. This is far from clear. It certainly wasn't considered as such by anybody until the 20th century.
My thinking was more so that, regardless of legal enshrinement, “freedom” is a pretty universal component of the American ethos. (as I understand it)
You’d imagine that politicians would love to use people’s emotions on the topic, regardless if they have a de jure reason to oppose a change like gun control, or not, in the case of cars with remote off switches.
But I'll leave that there. I wasn't commenting to put forth a particular political agenda, I was just trying to inject a counterpoint to a thing that I've been hearing more and more that I believe is factually incorrect.
Actual text from the 2nd amendment—pretty obvious why your comment is being flagged.
e.g. https://www.onstar.com/services/stolen-vehicle-assistance
A killswitch that GM puts on your car without telling you because some marketer thought the value-add was a good strategy is not much different than a killswitch that GM puts on your car without telling you because someone else told them to do it.
In both scenarios, a third party can control your car without any real affirmative consent.
If it's now included by default on every car, then yes, I agree with you.
Either way, new cars will factually not have kill switches by 2026 because nothing will be defined until then. How long after is an unknown.
Edit: I made an error with my dates. Lawmakers have until Dec 2024 to make the rules. The earliest a manufacturer can enact the rule is after Dec 2026. But lawmakers can also extend the dates and also have until Dec 2031 to submit barriers and reasons why the system may not work.
https://www.congress.gov/bill/117th-congress/house-bill/3684... SEC. 24220. ADVANCED IMPAIRED DRIVING TECHNOLOGY.
And yes, it's prone to all kinds of abuse by all kinds of "interest groups" …
And it never really does.
Obviously I want nobody hurt. But the boomer and silent generation house and senate have, lets just say, tech comprehension problems. I think a 10 minute demonstration to this issue is enough to get them to reconsider that their ideas are utter shit. Especially so if something inconveniences or otherwise annoys them :)
That’s horrifying if so.
In the USA this change to electric parking brakes was hastened by most people just not using the parking brake at all, instead just putting the car into "park" on the transmission. Many experienced drivers I know in USA do not know what a parking brake is or how to use it.
Obviously my experience is limited but genuinely curious - what problems have we suffered because people w automatic transmissions just use "park"? If "none" then I guess I can't lament the disappearance of the handbrake although I did find it odd that my highlander only has the button which engages itself.
When relying on the transmission for park, this engineering trade off results in the entire lateral force of the car resting on a tiny little piece of metal inside the transmission, that isn't designed to be constantly worn, the only thing keeping the car in place, easily replaced, etc.
There's also this aspect of - transmissions are mutable, they can go in a lot of different vehicles. You can have the same transmission in a 2000lb sports car and a 8000lb truck. How big should the parking pawl in that transmission be? How do you make that decision before you even know what vehicles the transmission will go into? What happens when your transmission that was originally designed for a 2000lb sports car ends up 20 years later in a 9,000lb truck towing a 3,500lb trailer? The parking pawl does not change.
Brakes will be serviced dozens of times over the lifetime of a vehicle, while the transmission might never be serviced. That's enough reason alone to leave the important task of immobilizing the car to the brakes, not the transmission, in my opinion.
Do we have problems (eg transmissions failing faster) because of this. Or people getting injured etc? Because if not, it's just a distinction without a difference?
EDIT: I think you added the first sentence of your post after I read and replied to it. That was the answer I am looking for, thank you.
* Apply the brake (the pedal)
* Place the transmission into park
* While the brake pedal is still pressed, apply the handbrake
* Release the brake pedal.
This means the car is held in place by the handbrake with the transmission park as a backup, rather than the transmission park being the only thing holding the car.
Bonus: if you park on a slope, turn the wheels towards the pavement so that if the car starts moving, it will hit the pavement edge instead of running straight down the road.
> Constant use of only the parking pawl, especially when parking on a steep incline, means that driveline components, and transmission internals, are kept constantly under stress, and can cause wear and eventual failure of the parking pawl or transmission linkage.
https://en.wikipedia.org/wiki/Parking_pawl#:~:text=Constant%....
If it's just for maintenance, I don't see any problem with it being electric.
Super common in some areas to see handbrakes used all the time (ex: Atlanta, very hilly, lots of names like "Druid hills" and "piedmont").
Super common in some areas to see handbrake usage discouraged entirely (ex: Northern states, where the concern is that it'll freeze up)
Super common in some areas to see no one care at all because it's just flat everywhere (ex: texas)
Anyways, other reasons for electric parking brake than already discussed is hill and stop light hold. Electronic brake pedal control relies on stored negative pressure on brake booster and cannot maintain pressure indefinitely, as opposed to electric parking brake which can.
If the normal brakes don't work (which, admittedly, is much rarer these days), it's your last line of defense.
It's less important than it used to be, with dual redundancy required by regulation and with digital systems monitoring both braking loops. But there's still single points of failure on the brake pedal (like the pedal itself).
It was? Who decided this? It wasn't SAE.
Is there another type of e-brake besides line locks and cables on regular cars? I’ve only messed with a few cars with electronic e-brakes and most everything else I’m experienced with is pre 2005. Have they cheaped out on the cables now?
e.g. https://data.embeddedcomputing.com/uploads/resize/1256/756/e...
And FWIW, a quick image search for 'electronic parking brake' only shows this style.
though with electric brakes there's no longer a long wire cable that can break (happens in the rust belt)
Also, pulling the electronic-brake in an emergency won't lock the back wheels like a handbrake can. The car will come to a controlled stop
Just imagine people accustomed to highly assisted driving in a metropolitan area suddenly forced to take over when two cars cannot manage to pass each other on such roads.
So I liked to watch "There's no such thing as a Self Driving Car - A MOTOMANTV RANT" by a car journalist here: https://youtu.be/8NJQCyv6BSk
Whether they can feasibly scale out to many other cities in an economic fashion is still unknown, but on a technical level they're definitely possible, because we already have them.
And yeah, cities and countries vary in difficulty, though as far as the US goes, SF is quite technically challenging in terms of urban obstacles (though the weather is of course quite mild).
And when I see how most modern cars handle signs like a speed limit to e.g. 30km/h with both additional texts like "from 9-15” and additionally on "monday-friday" or limited "for 300m" I always shake my head.
There's no human driver in the car. At most, the cars can stop/pull over and ask for remote navigation from humans (but not remote operation).
The hardware is more expensive than regular cabs. The software has cost billions, an may cost billions more.
Operationally, energy use should be marginally higher. Hourly cab-driver wages: gone. Insurance and write off — we'll see, but if the software is any good this should be considerably lower.
I guess it comes down to the big unknown: will it work at scale. Can we find enough cities with great weather, where enough people will use it? What will happen when the price goes up? What is a fair, sustainable price? Will that be higher than human-driven car?
> I guess it comes down to the big unknown: will it work at scale. Can we find enough cities with great weather, where enough people will use it? What will happen when the price goes up? What is a fair, sustainable price? Will that be higher than human-driven car?
If we assume the technology won’t improve any longer, this is true. But I don’t think that is a very safe assumption. Everything seems to point to the question being when rather than if.
People said the same shit about Phoenix: "ohhhhhh, but what about the cities that don't have extremely wide streets and almost no pedestrians???" Now that they're in SF, the goalposts immediately move.
Weather will be gradually solved, just like other technical aspects.
We have some interesting experiments (Cruise, Waymo) wandering around (a subset of) our streets, but they simply can't drive nearly as well as 98% of humans.
They achieve some measure of "safety" by failing safe e.g. stopping and putting hazard flashers on when confused. But they're not an effective or useful "service" for us, in the current state. Have you ever seen one try to deal with a contested left turn, or a 4 way stop with pedestrians constantly crossing?
Haven't tried it, but I hear Cruise is considerably worse.
> Have you ever seen one try to deal with a contested left turn, or a 4 way stop with pedestrians constantly crossing?
I've done considerable mileage in Waymo and these two situations it handles pretty well. I think the software is improving rapidly - I've reported some problems that happened consistently that a week later seem to have been resolved
The problem is by the letter of the law, the car can simply never proceed. People drivers don’t need to adhere to that letter so much, so can get creative, auto taxis can’t.
If you can agree that in 1000 years, it's extremely likely that we'll have self-driving cars, then to paraphrase Churchill, now we're just haggling over timeline. 500 years from now, will we have self-driving cars? 500 years ago, gunpowder and windmills were new technology. How about 200 years? The big thing was steam power then.
And so forth and so on. Unless there is some natural or physical law that would prevent self-driving cars, it's just a matter of when. And honestly, if you watch some of these Tesla FSD drives that people post on Twitter, it's pretty clear that self-driving is production-ready under a significant number of circumstances.
The problem is, of course, that there are a lot of edge cases, and because the stakes are high, we have to solve most of those to get true full self driving. But given that cars can pretty reliably traverse the streets of SF, I struggle to see the argument that the remaining edge cases are impossible to solve.
Ignoring the questions of if there will even be people -- let alone cars as we known them -- in 1,000 years, my answer to this is:
I have no idea. I don't see any special reason to suspect that such cars will exist in 1,000 years. I also see no special reason to suspect that they won't.
> If you can agree that in 1000 years
We don't agree on this.
At any time between now and 1000 years from now, will we have the capability to build a fully self-driving car, assuming that there are no cataclysmic, apocalyptical, or other type of events that render any question of technology irrelevant?
His question is - can your mind wrap around the amount of progress that can happen in 1000 years? If yes, can you imagine that going from "now" to "fully automated cars" is a small amount of progress, relative to progress that's possible on 1000 years? If yes, great - now you can dial back that it probably won't take 1000 years.
This will be the best of both worlds because we’ll still be able to drive when we want or need to (on roads where it’s allowed), and free to be driven when we prefer that, and the existing fleet of cars can age safely, preserving that investment for its useful lifetime.
It took a bit longer than a human driver would, but it made it through. I should have videoed it because I thought for certain it would get stuck.
My biggest complaint with Waymo's driving was it sometimes didn't plan ahead to be in the correct lane for a turn coming up within a few blocks and then couldn't get over due to aggro drivers. They seem to have addressed this though
The most impressive feat was when it went down a narrow road to find it was blocked. It managed a 3-point turn in a difficult situation and found another route.
The biggest "bug" I witnessed was a situation with only one road going to the destination and that road was closed with a police road block. It turned around and then kept coming back to the road block, basically looping forever. Had to have an engineer intervene.
If Waymo can handle SF roads, I think it can handle anything in the US (at least without deep snow on the roads). Dunno about other places where you might have to get "creative" with your driving.
How does one tell the robot car to pull over, I'm done and getting out here? Can you open the door and jump out if need be?
[citation needed]
This is a very dubious statement and it really sounds like you have never driven a car or anything larger.
Human perception + reaction time is eons compared to how quickly a computer can move the wheels and pedals.
Whether the machine makes the _right_ decision is much more difficult, but a computer can actuate with speed and precision humans simply cannot.
This is why I said what I said; the OP wording implies they meant just that.
I was frightened, but impressed. (Good thing nobody opened the door.)
re: 1.5 car width lanes or trails or whatever
Have you seen humans drive in other countries? Italy, India and Boston (lol - many more) have driving situations where lane lines are not obeyed and cars pack pretty tightly on a regular basis. I cannot imagine a world where that is not solved.
The issue are the millions of edge cases in which humans remain far better at analyzing, figuring out a solution, and compromising in.
Well, we are approaching physical limits of minimum transistor size, so that might not be as ridiculous as you think. What if the best we can do is to triple the current processing power, but that's not enough?
Here's an example of a non-autonomous car backing up for a FSD Tesla: (at 22:20) https://www.youtube.com/watch?v=2l6KLC3S5gc&t=22m14s
FSD Beta handles this now, including the negotiation of which car passes first
I soon will be on holiday in rural France again and you might not believe who and what is driving on those small roads. Which is totally OK with me, but I highly doubt tech will handle that in the near future.
Humans aren't really all that special.
As long as there is a person with access to that surveillance, it will be used to shame and control people. If England had the surveillance it has today 100 years ago, they would have ousted Alan Turing for being gay long before he would have had the chance to help them win the war.
Or remotely nerf your car if you miss a few payments?
It won't matter if you participate or not - the sensors are a miasma that surround you, and your behavior will be captured and inferred anyway.
The world of Enemy of the State (prescient film) is real [1].
Just wait until we get DNA sensors that capture the air and endothelial cells you exhale [2]. Those are coming too. Perhaps a lot like the ones in GATTACA.
[1] https://en.wikipedia.org/wiki/Enemy_of_the_State_(film)
[2] https://www.theguardian.com/environment/2023/jun/05/airborne...
No antenna. No signal. No datalink.
But yeah, no guarantees that this is possible, or terminating all external antenna connectors is sufficient.
I can put my car in offline mode (blue link).
I mean if everyone starts riding bikes they’ll just apply the same tech anyway.
There’s no winning this arms race, the only protection is going to be laws so we should really be putting much more care into laws, the people who make them, and the systems that put or remove them from power.
Gait is one of those modalities that gets rediscovered every decade only to, again, yield no results.
There's just too much noise and variability over time for it to qualify for broader use.
It's much easier to get a detailed image of a person's iris from 100m.
Seriously? That sounds insane, wow. I mean, an iris is like 1cm wide, can you really get a detailed image from 100m away, while the person is walking, looking around, blinking etc.?
EDIT: I mean in the context of surveillance, I can imagine that a dedicated photographer could get such a picture with a fancy camera, but we're talking about 24/7 video surveillance.
I assumed there would be an appropriate setup for that.
In any case systems collecting iris data en masse from 6m were available a decade ago. Ultimately it's just a photo in near infrared, nothing fancy.
The whole reason why this even works is that the iris as a modality has an unrivaled 240+ bits of entropy, so even partial or blurred images yield enough information to identify a person(requiring 60bits at most).
On top of that it's an extremely stable feature. Shabrat Gula was famously identified after almost 20 years, despite being a child when her photo was taken.
https://www.robots.ox.ac.uk/~az/lectures/est/iris.pdf
By far the largest database of irises is the one created by India’s Unique Identification Authority, with more than a billion people enrolled:
https://www.irisid.com/indian-national-identity-program-tops...
Surely the value of gait over iris would be that you don’t need ridiculously specialized sensors right? Can use (hypothetically, I guess) a bunch of CCTV sending that’s already out there.
Samsung used to put them in their phones and they seem to have been using just one frequency.
Even better: the wavelengths used are in the range of 700-900nm, which, as indicated by my smartphone camera picking up the 850nm diodes in my Leap Motion controller, are within the range of off-the-shelf camera sensors.
For a decent image you need the iris to span 70 pixels. A Full HD camera probably won't cut it, but a 4k one installed at the entrance of a building might well be scanning irises and we wouldn't even know it.
Although it isn't enforced, its already illegal in most places to wear masks in public though. I've seen exceptions carved out for children (because of Halloween) and for people with doctor's notes, but not for cyclists
*does not negate the other ways they can PiD you.
I think that's quite an important *
if you have a phone in your pocket, you're already trackable 24/7
Because of this, bicycles played a significant role in women’s liberation movements.
https://www.womenshistory.org/articles/pedaling-path-freedom
Sounds a bit improper but this is also how when a person robs someone they’re able to put out security footage pretty quickly: “helpful neighbors” actually being legitimately helpful neighbors.
To be clear, the surveillance cameras on legs are cheaper to query than the surveillance cameras on wheels
A network of spies is far more expensive to maintain than a network of cameras, wheeled or otherwise.
The claim was "We're prohibitively expensive to poll", and I am correct it is cheaper to poll humans. i.e. query X regarding Y, where X is {humans with eyes|cars with cameras}, and Y is the evaluated state of some location at some time.
I do agree if the poster had written "it's cheaper to set up a network of self-driving cars than spies"...well when I put it that way :P...but jk sure I'll say spies.
https://en.wikipedia.org/wiki/The_Entire_History_of_You
https://en.wikipedia.org/wiki/Crocodile_(Black_Mirror)
> an insurance investigator who uses a "Recaller" to view his memories, as best as he can picture them, on a screen
[1] https://www.reviewgeek.com/111381/you-dont-really-ever-own-a...
[2] The Hated One Your Car Is a Better Spy than Facebook https://www.youtube.com/watch?v=WX2SWUMt_fk
[3] https://hothardware.com/news/bidens-infrastructure-bill-mand...
[4] https://www.theguardian.com/technology/2023/may/26/tesla-dat...
[5] https://doubleagent.net/2023/05/21/a-car-battery-monitor-tra...
[6] https://www.wired.com/2015/07/hackers-remotely-kill-jeep-hig...
Frankly with higher interest rates and reduced investment, it’s easier to imagine research being scaled back than true self driving cars becoming generally available.
To put it another way, a computer that only the 5 richest kings in Europe can afford is still a computer.
With self-driving cars, I'd argue mixed bag. Yes, it's already possible (with caveats), but "no" it's also at least five years away. When can I have it? When would it make sense for an existing cab company to jump ship? I mean, fire the drivers and replace old fleet with commodity self driving cars.
Personally I'm not so interested in what kings, millionaires, governments can achieve unless it's either a stunning scientific achievement or somehow relevant improvement to society.
https://en.wikipedia.org/wiki/List_of_driver-less_train_syst...
But by and far the majority of trains are still human operated. Even though the traffic rules are very straight forward ("get out of my way, or...") and the guidance systems, robotics, CC surveillance, scheduling and logistics are 99% there in (for example) major parts of continental Europe.
So why are train services in Germany, Netherlands, France frequently canceled for lack of train drivers? I honoustly don't know the answer but I wager that return on investment plays a significant role. And I think the self-driving car rationale isn't all that different, except for the easy venture billions of the last decade(s).
The reality is it’s political suicide to put that many people out of work in one fell swoop and it’s expensive to replace all rolling stock in one go.
So it’s down to a slow and steady war of attrition.
Example: London Underground is being slowly but surely automated, while keeping an operator who is “not a driver” until they are no longer needed. And that’s one of the oldest metro train systems in the world
So, were the train experiments successful? If so, what is stopping it from replacing regular trains?
From a distance, trains should be simpler: it's just a horizontal elevator. So why are there still mostly regular trains? I think answering that will likely help explain why self-driving cars won't be a thing in at least another 10-20 years.
> And it’s easier for law enforcement to turn to one company with a large repository of videos and a dedicated response team than to reach out to all the businesses in a neighborhood with security systems.
And both of those are much easier than identifying you and reaching out to you in order to pull the recordings off of your Subaru.
What if government agencies and oligarchies benefit form that stalking, too?
(replying to what you didn't say) Extend the existing dumb automatic copyright to our personal data so that buyers/sellers have to pay royalties.
Have you seen the terms of service for social media sites? There's a clause in there that grants them a non-exclusive transferable irrecoverable royalty-free license to whatever you post on the site. What makes you think that companies won't add the same clause for your personal data?
What else needs to be uncloaked as surveillance technology before people stop trusting big tech?
Whatever your opinion of him, Snowden never would have made it out of the country. His own car would have detained him.
Stay inside.