Who's Responsible For Accidents Caused By Open Source Self-Driving Car Software? (ieee.org)
Here's the problem. "You could download Comma.ai's new open-source Python code from Github, grab the necessary hardware, and follow the company's instructions to add semi-autonomous capabilities to specific Acura and Honda model cars (with more vehicles to follow)," writes IEEE Spectrum. But then who's legally responsible if there's an accident?
Long-time Slashdot reader Registered Coward v2 writes:
While many legal experts agree OSS is "buyer beware" and that Comma.ai and its CEO Georg Hotz would not be liable, it's a gray area in the law. The software is release under the MIT OSS license and the Read Me contains the disclaimer "This is alpha-quality software for research purposes only... You are responsible for complying with local laws and regulatons." The U.S. Supreme Court, in a series of court cases in the 1990s, ruled open source code as free speech protected under the First Amendment of the U.S. Constitution.
The question is does that release the author(s) from liability. The EU has no EU wide rules on liability in such cases. One open question is even if the person who used the software could not sue, a third party injured by it might be able to since they are not a party to the license agreement.
An EFF attorney told HotHardware "Prosecutors and plaintiffs often urge courts to disregard traditional First Amendment protections in the case of software." But not everyone agrees. "Most legal experts that spoke with IEEE Spectrum -- and Hotz himself -- believe that if you use the company's code and something goes wrong, then it isn't liable for damages. You are."
The question is does that release the author(s) from liability. The EU has no EU wide rules on liability in such cases. One open question is even if the person who used the software could not sue, a third party injured by it might be able to since they are not a party to the license agreement.
An EFF attorney told HotHardware "Prosecutors and plaintiffs often urge courts to disregard traditional First Amendment protections in the case of software." But not everyone agrees. "Most legal experts that spoke with IEEE Spectrum -- and Hotz himself -- believe that if you use the company's code and something goes wrong, then it isn't liable for damages. You are."
Software liability is not new. Its use in automobiles is.
If you download, compile (or trust some 3rd party build), and use you are responsible.
If your self driving car, using open source software, is in an accident, you are responsible. There is no responsible 3rd party.
This is old law.
This is the meaning of use at your own risk.
Accept it, or don't use it.
Sure, the store may go after Joe the installer, but that is their problem, not yours and the reason you didn't hire Joe directly. See subrogation.
I have to agree here.
If I buy a part for my car and the part's manufacturer claims that it complies with some ASE or similar standard, then if the part fails I might have a legal case (e.g., if I can prove negligence in the design or manufacture, or something that the established case law will respect). However, if I buy a part off a guy who makes them in his tool shed and hey tells me "hey, I'm not sure that this thing won't explode when apply the brake," then I am pretty sure I have no recourse whatsoever.
How is software different? If the manufacturer warrants it, then it should work as it is represented and if it fails then there is a discussion to be had. If the manufacturer disclaims warranty and it breaks (and the applicable laws don't override that; you know that in some jurisdictions that there are laws that still hold the maker or seller responsible to a degree for things they make or sell?) then I don't have a legal case.
Of course, even professionally produced commercial software normally has a EULA with a clause that reads something like "the manufacturer provides no warranty of merchantability or fitness for a particular purpose and shall not be liable for losses arising from blah, blah, blah..." If you are NASA and paying your contractors (an enormous amount of money) to mathematically prove their software correct then you might get a "yup, we certify that this software will work as designed," otherwise you have no such assurance.
It's easy with systemd. You just have to write it in the form of a Sanskrit poem.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
Actually, It's me. Sorry in advance for all the dead people, all of that is my fault. Even the ones ain't died quite yet. Although I do feel badly, I do. Again, sorry.
This is one of my arguments about so-called 'self driving cars': The manufacturer is technically driving it if it's in self-driving mode, but just like so many of them, they'll dodge responsibility one way or the other if someone gets killed, and in the end there'll be no justice at all. It'll get tied up in the civil court system for years, or in some sort of arbitration, and in the end you'll either get nothing or some paltry chunk of change, and meanwhile a person is dead, all because of some shitty technology that was rushed to market that doesn't even have the cognitive ability of a smart dog. These so-called 'autonomous/self driving cars' are not going to be the panacea that some of you think it's going to be, we're just going to trade human error behind the wheel for human error in the development division of an auto manufacturer, or human error in the marketing department for rushing it to market before it was really safe to do so. All I can say is I'm not going to be the one strapped into a car seat with no controls whatsoever to stop the gods-be-damned thing when it goes haywire and kills me, and I have nothing but feelings of horror for whoever does, and deep sympathy for whoever the poor bugger leaves behind them.
Oh, and by the way? The way I think this should be handled legally-speaking, since it's obvious we'll be subjected to these gods-be-damned things regardless, is the same way aircraft mechanics are treated, legally: If a plane they worked on crashes and people die, and the cause of the crash is proven to be a mechanical failure that's the responsibility of the mechanic that worked on it, he is arrested and tried for murder. A so-called 'self-driving car' kills someone? The programmer(s) responsible for not doing their job correctly get thrown in jail charged with murder. Oh and before any of you give me shit for this? You all make a big point about how 'self driving cars will save lives'; well if they TAKE a life then some HUMAN has to be held criminally responsible for it, plain and simple. Otherwise you're just hypocrites.
The owner of the vehicle is responsible for ensuring his vehicle is safe to use. If he modifies it, by installing some untested software, he is most certainly responsible for the consequences if he then injures someone. And if the disclaimers are clear enough, his chances of successfully suing the software developer are slim.
The sad thing is that, as with the Tesla, you know some idiots will install this and go on to kill people - hopefully just themselves. After that I wouldn't be surprised to see the software developer sued claiming the warnings on the software were just not clear enough. Or even the car manufacturer, for allowing the vehicle to be so modified in the first place...
> The reason we have shitty software is because nobody typically dies when something goes wrong
The reason we have shitty software is because nobody wants to pay for having well tested software. The costs would be factors higher.