IMO the title is a little misleading - I read 'EU drops rule' as that the EU had removed a rule / cancelled it.
Should probably be something like "EU states that telemetry data collected from self-driving cars may be subject to copyright protection".
Then the bigger discussion should probably be - 'What is telemetry data?'.
If telemetry data is _any_ data collected from a car that could be remotely accessed, then in theory I'd argue that digital oil sensor readings, digital odometer readings and digital tyre pressure readings are all telemetry data if they are stored in a place that can be remotely read.
If that's the case - it could lead to the situation where car manufacturers demand (by law) that you are only allowed to service your car with them and not be a mechanic of your choice.
That may start as a small inconvenience that luxury car owners may be able to afford but as those features come to lower end models it may hurt average or less well off consumers.
Further more, what happened to 'facts' not being copyrightable (generally speaking).
If I take a picture with my camera the camera manufacturer will not get the copyright. How on earth are car manufacturers gonna be the ones that get the copyright?
I don't think this is a fair comparison. It's like saying all the work I do on my computer belongs to apple. Which doesn't make sense. But apple does keep logs and telemetry of performance and crashes and uses that to improve it's products.
With self driving cars, and other devices (like apple electronics) the logs they keep give you really really good insight into what sensors, techniques, and algorithmic approaches they are taking to problems, and those directly expose proprietary information. I would argue that this information entirely belongs to those companies (with an agreement to send this data to their servers).
Also most AI and self driving companies use the data collected by their existing fleet to further improve AI systems in these vehicles. The competitive advantage of most of these companies is how much data, with what sensors they are able to collect. Making that information public will have a huge negative impact on the tech and AI industry, essentially pulling the carpet out from underneath everyone.
> But apple does keep logs and telemetry of performance and crashes and uses that to improve it's products.
Sure, but do they claim copyright on that? Am I not allowed to paste a log online?
... and those directly expose proprietary information.
How would copyrighting that information make any difference? As long as competitors can get to your logs (which is orthogonal to whether they hold a copyright) the copyright doesn't stop them from gaining that understanding. They could even publicly disclose their understanding.
Just encrypt the logs. Then no one will get at them and even if they manage to the act of decrypting them is illegal in most places so they probably won't publish it. And even if they do that it will be from one car, not a whole fleet.
I doubt that public telemetry would have a negative impact on the tech itself. Quite the contrary. Not necessarily true for the business case of course. And although there is a connection, it would boost the tech if there would be publicly available telemetry.
But your camera manufacturer may log on a dedicate flash memory inside the camera some logs about pixel defects, and claim that they copyright these data.
First, this analogy is broken, in that the MPEG LA is not the manufacturer of your camera.
Second, you may still use that footage in any commercial manner, royality free. At worst, you need to transcode it into any open standard. The footage is always yours, only the encoding might be encumbered.
MPEG LA won't get copyright on my material. It's just that the format used is patent encumbered. I guess I could just transcode it to a different format and be safe?
If you buy a used computer, the windows license doesnt transfer with it.
Ignoring that however, you cannot be bound by a EULA that you receive _after_ signing a contract. If you receive the eula before or with a contract, it may be valid
I don't think Windows EULA apply to you even if you buy a new one. You didn't sign anything. In fact, maybe it was your friend helping you to setup your computer that pressed "accept".
In EU pressing accept is legally binding, along with other forms of agreement like gentlemen (handshake) agreement, over phone, etc. I can open a bank account by pressing accept.
Source? I certainly need to digitally sign something like that. Also opening a bank account might not even require something legally binding. Just that the bank will refuse service if you don't obey their rules. (and if you abuse it you are most certainly committing fraud anyway)
In any event it doesn't apply here. Because noone can tell who pressed accept.
edit: Also, if you open a new account as an existing customer that's because you've agreed to the terms previously and identified yourself by logging in.
I wasn't talking about being an existing customer. This very month I opened a bank account in another EU country using a scan of my EU ID and driving license and then applied for a 20k EUR loan, without digitally signing (or normal signing) anything, their nearest branch is around 1k km far. Just a video phone call where I shown them my ID next to my face, the scans, and a code that I got over sms and entered into their website. The banking regulation requires a legally binding contract for all banking services. There is a loan service called Zonky that literally doesn't have a physical branch, it's all online, no digital signing as well; it's a P2P service (I used to invest there) and the people using it have zero idea what a digital signature is.
I'm not saying you're wrong, but this explanation seems weird to me. If I get someone to accept instead of me, can I then abuse the software legally and they can't even (legally) revoke the license? That is very weird. What about Netflix accounts, is my girlfriend not bound by the ToS when I was the one who clicked Agree, and thus she could legally rip their content and share it? Very weird.
You did provide ID which I believe is a big difference compared to just pressing accept. No signature?
Just because the EULA isn't legally binding doesn't mean that the company in question can't act upon it. But you have no legal obligation to follow the EULA (because you didn't sign it).
When it comes to a "physical" license that makes it a bit more tricky, I'm not sure of the details. But they certainly can revoke the online-aspect of a product (tesla will (I believe) revoke super-charger access and mobile-app access if they feel that you violated their terms).
If you cheat in a multiplayer game that will probably violate the EULA (but not any laws). If you are caught you will likely be banned.
> * What about Netflix accounts, is my girlfriend not bound by the ToS when I was the one who clicked Agree, and thus she could legally rip their content and share it? Very weird. *
You both could legally rip their content (as long as you don't break any laws doing so (and by ripping and circumventing their encryption you most likely would break a few laws)). If they detect it they can and probably will terminate your account (even if what you did was legal). But you didn't break any laws by not following the EULA.
Your friend, or the new used car owner, will get to push a button every time the car starts, agreeing to the contract. Yeah, yeah, unenforcable, blah, blah. You want to go to the mall, or not?
If we have EU cookie banners every time we turn the key in the ignition, I give up. I'm going to find some 80s pickups in Arizona or something and start stockpiling them.
Yup, there's a "decline" button. The ironic thing is that my acceptance or declination doesn't go anywhere. The original Leafs used frequencies (GPRS) that AT&T quit supporting. Since the now-worthless radio is a useless threat surface, I had Nissan remove it for free. Because CarWings is a mere half-assed shell of what it could be, I declined to spend the $200 on an updated cell radio. So if I do press a button, it's for naught.
Anyway, all the screen says, in essence, is that Nissan is using the aggregate data blah, blah, and if you want to use the nifty history telemetry, well, you're going to have to let the car send those data. Or not; car works just fine if you press "decline".
Sure you did, a contract of sale. That makes you the owner of the device. Should another person be able to access all your computers sensor data just because you let the person use your computer?
Of course my friend can use the sensor data of my car if I let him/her. The question is whether the car company can claim that the sensor data that my friend collects while driving is theirs without informing him/her of it.
>"That may start as a small inconvenience that luxury car owners may be able to afford but as those features come to lower end models it may hurt average or less well off consumers."
Longer-term, I wonder if that isn't what will ultimately happen anyway if the concept of car ownership sufficiently changes to be predominantly renting from a pool of vehicles, or a model where you "own" it, but lease it out to the pool.
Whether you can get a dump under the GDPR is separate from who holds the copyright, and even if you can, I think a strong argument can be made that not all data that such a car collects need be personal data. Where your car was, what passengers the car detected, whether they were attentive, etc, to me certainly are personal data, but a log of, say, oil temperature or vibration in the engine may not be personal data, certainly not if the manufacturer immediately aggregates it across thousands of cars.
Also, this case doesn’t (as of yet) declare that data is copyrightable; it just removes a clause that explicitly states it isn’t.
You can extract out a lot of information if you want to though. Perhaps vibration + temperature over time leaks information about where you are: gravel road or not, pot hole or not, hit the curb pulling into your driveway vs your friends driveway, etc. Its hard to not leak data now that machine learning can look for the patterns.
Hypothetically I think it may make sense to limit arbitrary mechanic shops from repairing driverless cars. If they mess something up that causes the software to malfunction, then the fault of blame gets even more obscure. I would instead hope that such a change would be coupled with reduced car ownership.
You're making the assumption that the dealerships/manufacturers' mechanics are more competent than independent mechanics. I don't believe that is necessarily the case.
The title is very misleading. The parliament just voted to not include a clause that would have explicitely made that data not copyrightable. Unfortunately, the article does not tell us why this clause was removed. Maybe it was just removed because it is unnecessary?
Automotive telemetry data in many countries is subject to multiple regulatory regimes, not the least of which is GDPR, related to use of this data and who is entitled to access it. Exclusion from copyright protection would appear to be at odds with these regulatory regimes. Entitlement to access usually includes the government and manufacturer.
How so? The scope of GPDR includes a lot of data that is not copyrighted or copyrightable – including anything purely factual in nature rather than creative, such as my IP address, net worth, choice of religion, etc. (Well, technically in the EU you can own 'database rights', which are similar to copyright, on large databases of facts, but only in certain conditions and subject to expiry and exemptions; GPDR has no such limits.)
This data is the equivalent of tracking all the data off the mobile phone of an unknown person -- it is anonymous but trivially identifiable. The data is also definitely copyrightable in nature, being complex engineering analytics customized for the vehicle by the manufacturer and used for a wide variety of internal purposes. The contents and semantics of this data varies widely across manufacturers. It goes far beyond dumb telemetry.
Much of this data is treated as a legitimate trade secret by the manufacturer and is generally deemed to be owned by the manufacturer for legal purposes, though they are sometimes required to share it with regulatory bodies. Automotive companies would strongly object to this data becoming public because it would adversely impact their ability to maintain a technological competitive advantage. That their engineering trade secrets are covered under GDPR makes it that much more awkward, but owning a copyright on this data at least gives them legal standing to enforce compliance. I've worked with major European automotive OEMs -- they would lose their minds if you told them this data is no longer theirs to control.
Likely somebody paid for it. EU is creating the law for big corporations, so it seems like one of those would like to copyright telemetry data and EU just amended it to please them.
They are the centre-right coalition of all centre-right political parties in Europe. It is beyond laughable levels how they corrupt they are. You can predict their stance 100% of the time by assuming it is the business friendly position.
Any country with "democratic" or "people's" in the name is not democratically run and the people have pretty much no say in anything (Congo, DPRK, etc).
Any armed faction that references a religious figure or text in their name commits atrocities condemned by that religion (LRA, ISIS, etc).
Any state that calls itself "commonwealth" is run by a political elite that does a piss-poor job looking out for the interests of the commoners (maybe worth granting an exception for Kentucky, I've never been there and don't know many people from there so it might be alright).
The car dealership that has radio ads talking about what a great value they are is nothing of the sort.
A political party that prioritizes business over people calling itself the "people's whatever" seems pretty standard as far as naming conventions go.
How the hell did North Korea get the title "Democratic People's Republic"?
Edit: To be clear, I'm not trying to compare the People's Party to the North Korean regime, just pointing out that organisations can pick their own name, and that doesn't necessarily have to reflect what they are, or what they will become years after they have chosen the name.
It's a direct translation of the name of their founding member, which is the Spanish Partido Popular. And yes, they're right-wing corporatist authoritarian in Spain too.
Your comment is a bit weak ("can we have a rule") - but this is exactly the type of thing we should be working towards. Why should software vendors gain legal control over our data, simply because we use their software?
If the compiler vendors could have pulled this crap back in the 80s, the Internet as we know it would not exist.
Saying "a rule" sounds like powerlessly wishing a problem gets fixed. The term makes sense for the OP about a detail in the context of a larger process, but you're talking about a more general topic that needs a wider legal-philosophical push with how metastasized copyright and adhesion contracts have become.
My question was rhetorical - asking what is right, not what legal mechanisms are being abused to create the current state.
> Saying "a rule" sounds like powerlessly wishing a problem gets fixed.
Isn't that mostly the case when we post about some (perceived) injustice on HN?
Take for instance the patent system. Every now and then people come up with ideas for what the system should really be like, but nobody ever provides any hints for actually getting these ideas implemented, and almost always the forces that keep the current system in place are totally ignored.
> Isn't that mostly the case when we post about some (perceived) injustice on HN?
Talking generally about "what ought" seems distinct from powerlessness. At the very least it delegitimizes the current status quo, even if only in a subculture. For instance, all of the HN (and wider tech) discussion of the NSA has done nothing to actually get those criminals any closer to prison. But designers of new protocols start with a background of making them surveillance-resistant rather than surveillance-supporting.
I apologize for starting my original comment somewhat aggressively - I wanted to draw attention to your idea behind the phrasing, with constructive criticism addressing why (hopefully) yours was downvoted, but edited it to be (too) simpler.
This is a bit radical but here is a question - what is the public interest that copyrighting this data serves whatsoever - regardless of who it belongs to?
Isn't all telemetry data currently copyrighted? Streetview? Pictures of a street? Satellite images?
Why would it be public domain?
Honest question, I'm not understanding how telemetry data is different than airborne laser scanner data, Streetview, map data, satellite pictures, etc.
Or even server logs generated by an iPhone for Apple, Tesla's car data logs, etc.
There's some precedent in the US that "factual" photographs that have no creative element are not copyrightable. In practice, I think various contracts and EULAs will override copyright questions anyway.
Somehow I have thought that monopolies are so generally accepted as harmful, that it would be the proponents of monopolies that have the burden of proof. Not the ones who oppose monopolies.
So instead of just assuming that copyrights are good for society, can you provide some evidence?
Copyright of a picture of a street isn't a monopoly on that street, you can head over there and take a picture of the street yourself.
Other than that, just thing of photographers who spend their life working on photography. Copyright allows them to charge for their work, or for a software developer to charge for their work.
Removing copyright entirely means you can only monetize your work based on the marginal cost, the result is zero monetization for any type of work that has near-zero marginal cost, such as a book, movies, photography, etc.
Sure, you can argue that we should move to the fairyland dream scenario where everyone is beautifully funded by fans and so on. We've been there, it was how art in the Middle Ages worked.
Anyway, that is a separate discussion:
1) Given that copyright exists, should X be copyrightable?
Sure, but you're also party to an adhesion contract where you're forced to surrender that ownership to the manufacturer. As Coase pointed out, the initial allocation doesn't matter so much as the market will get it to whomever can best exploit it. The liability is in carving out a market-legible ownership handle to begin with.
But copyright is likely moot anyway, as the cars will be imprisoned [0] with strong digital restrictions to siphon the data to manufacturers while preventing access by people.
EU laws in general are quite hostile to adhesion contracts - pretty much all terms that where a consumer is forced to surrender something meaningful to use a product they bought are considered null and void in EU. If customer is shown a common EULA stating "you surrender copyright of these things to the manufacturer" and needs to click "I agree" before proceeding, then in EU that does not imply any legally binding commitment whatsoever and does not imply that the consumer surrendered those rights; if they were forced to do so in a standartized mandatory agreement, then that doesn't count.
Doesn't Feist preclude copyrighting raw data that wasn't created through some creative spark? That's in the US, granted, but I'd hope similar principles would apply in other jurisdictions.
Quite the opposite. Telemetry data is an intended product, because it allows for improvements to the underlying product and much more (such as mapping, for example).
Think of it as Google Analytics. If you run a site and use Google Analytics on it, do you own the copyright of the data generated by it?
Telemetry is simply gathering of factual information. And it's not produced by a human. There is no copyright for something produced by machines or animals, so far.
You can consider copyrightability of something produced by intelligent machine (human like), but that's a whole topic in itself and not even the case here.
A map is a dataset of facts. The MLB stats are a dataset of facts.
Copyright doesn't mean that you can't create a work that is similar based on the same facts, it just means that to do so you can't copy the original work.
You can create a new map of the streets, but you can't just do so by copying an existing map (with the exception of certain licenses and so on).
Copyright is supposed to protect creative works. There is such a thing as a threshold of originality. There are database rights, but that is separate from copyright.
I can't wait to cite this when another person claims that e.g. the GDPR is so anti-corporate, and the EU's only goal is to punish American companies.
Somehow the EU is always the bad actor, it doesn't matter which side it ends up favoring in some specific action.
Of which there are probably thousands, if not tens of thousands every year. But sure, the world ending always hinges on exactly the one action some commenter happened to dislike that day.
In general, or in this specific case? Because I can think of a number of times us vs. them have been important for the advancement of justice, civil rights, protection of individuals etc.
Copyright is intended for the protection of literary or artistic works, but it's been extended to cover recordings and performances. Those arguably have some artistic or creative properties, but I find it hard to argue that telemetry data has any real creative properties.
The truth is, you don't need copyright on telemetry data for machines you control and own, because that data never leaves your control.
It only matters for machines you don't both control and own. Who benefits from differentiating between ownership and control?
This seems to open a big can of worms: If it can be copyrighted, it has an author. So who is it?
* Is the driver the author of this data? If yes, does the car company need to get a license to fetch this data or are they pirates? Is the driver permitted to exercises authorship rights like distribution? Why not?
* Or is the car's creator author? Then the driver had no influence on the creation of the data, so can't be responsible for anything: His actions had no consequences.
* Or are both parties co authors? Then you need agreement from every author before you can do anything. This causes both problems at the top at the same time.
I'm not sure how we're defining an old car here, but I'm pretty sure a car from 2010 is not noticeably more dangerous than cars today.
Preferring a car with more freedomey™ qualities because you don't want to be tracked, analyzed and want to increase the security of your car (as in, to be managed only by yourself) hardly qualifies as being a luddite.
I'm going to go out on a limb and suggest very few people will own an autonomous vehicle and will probably rent/lease them instead, which solves the problem neatly, if it happens.
If you don't actually own the thing, then it's harder to claim the data is yours.
Neatly for manufacturers. If the users don't own the vehicle, the data remains the property of the manufacturers or the business renting them to users (though I'd bet most of it would belong to the manufacturers).
Should probably be something like "EU states that telemetry data collected from self-driving cars may be subject to copyright protection".
Then the bigger discussion should probably be - 'What is telemetry data?'.
If telemetry data is _any_ data collected from a car that could be remotely accessed, then in theory I'd argue that digital oil sensor readings, digital odometer readings and digital tyre pressure readings are all telemetry data if they are stored in a place that can be remotely read.
If that's the case - it could lead to the situation where car manufacturers demand (by law) that you are only allowed to service your car with them and not be a mechanic of your choice.
That may start as a small inconvenience that luxury car owners may be able to afford but as those features come to lower end models it may hurt average or less well off consumers.
Further more, what happened to 'facts' not being copyrightable (generally speaking).