CafeNervosa
16479
509
11
Sauce: https://www.wired.com/story/openai-backs-bill-exempt-ai-firms-model-harm-lawsuits/
Apr 10, 2026 10:58 PM
CafeNervosa
16479
509
11
Sauce: https://www.wired.com/story/openai-backs-bill-exempt-ai-firms-model-harm-lawsuits/
thisuserisafk
Yeah. Saw this coming when we used AI to bomb a girls school.
QuatermainFT
this is why asimov thought ai tasked with protecting us would decide we needed to be rounded up and kept in pens to protect us.
tallroses
Laugh all you want, but AI companies have REPEATEDLY refused the military’s demands that AI be allowed to “pull the trigger” and the government has not taken no for an answer. They have no choice at this point than to protect themselves legally or they will definitely be the fall guy in the end.
BuckinCrazy
Well fuck
jalcantara88127001
Fuck al the way off. EthnicCleansing2026.exe
Ankylosaur
I'd back a bill that would limit liability for people who throw Molotov cocktails at Sam Altman's house.
Randomice
Was that bill written by AI?
Sebastopol140
Of course.
CafeNervosa
Wouldn't surprise me if it was.
PowerThruster
Well of course theyd need a bill like this before rolling out automated AI weaponry. Can't be liable when the newest Boston Dynamics shoots up an entire building chasing a suspect for jaywalking or littering.
VagrantKoba
At the headline I thought they meant "you can't blame the program after a certain point " like sure if AI accidentally sends money to the wrong account, that's a whoopsie on the program's account, fix it move on, but if it sunk everybody's money into something we'll then the company is liable, but no. It's worse. -sigh-
CookieMonstersCrumbs
They kn ow the big crash is coming and this is what they're trying to do to save their own fucking ass
5kergi0
That headline, really? I'd like to know who in congress they paid off to write and introduce this bill because they need to be put on that list we're all keeping of Nazis that are REALLY bad for our health.
Kyzyl
JackieTreehornProductions
There's a former girl's school in Iran that is a prime example why.
Sticklebrickk
Google's "Don't be evil" comes to mind
AriaYuki
Shame they removed that from their policies
mafiacarstarter
I wouldn't be surprised if someone decided to try burning their CEO's house down.
galhnios84
As it should. If you trust a computer to kill people for you.. YOU should be held accountable.
EnderKing567
Rules for thee but not for me.... Hope they end up on the short end of the stick instead... >:(
jalcantara88127001
Time to steal robots and reprogram them with AI trained to kill CEOs. Win win
unluckyandbored
Of course they are.
hereforthememes420
W.I.T.A.F.
Rathan5
Well, it’s not TECHNICALLY “AI before humans”, it’s just “humans who own the AI companies before other humans”. Don’t forget, much like corporations, “AI” are not separate existences independently making these business decisions. They are being made by people, and are almost always made to either gather more wealth to themselves, or to shield them from being held accountable for the awful things they do in the pursuit of said wealth
jackal12345
Good news is openai is hemorrhaging money and will probably go under soon. Bad news is they're gonna do a ton of damage first.
Grimlinnan
CEO’s need to die for their crimes.
UncleRat
Hey, when 1.5 trillion is invested in something, and it still hasn't made anything like a profit, you're looking at WorldCom or Enron, not Skynet.
KenjiKando
Boy, so much for "no laws about ai for 10 years" huh.
justadude41123
Well when it comes to money, there is always a winner and a loser. Every financial disaster, the winner was never the meek…. The same when it comes to lives.. with the only small exception when they wish to tour the site of a great hubris, the titanic.
INeverReadTheTOS
"Corporation backs bill that would give it, specifically, legal immunity to do crimes" NO FUCKIN' SHIT SHERLOCK.
CeoHuntingSeason
YEAH! NOBODY SHOULD POST THIS "NEWS", ITS JUST SO FUCKING OBVIOUS THAT MAKES NO FUCKING SENSE TO TALK ABOUT IT, THEY ARE ALWAYS GOING TO DO THINGS TO GET AWAY WITH ANYTHING SO WHY BOTHER POSTING ABOUT IT? WATER IS WET, SCUM ARE DOING SCUMMY THINGS! DUH!
crodrigues
if this goes through then anyone can use AI to commit harmful crimes and not be held liable for it. I do not think these AI CEOs thought this through after what happened last year.
rustyroo1
There’s a hell of a lot of things these fuckers didn’t think about which have yet to manifest. It’s referred to as making boat loads of money and to hell with the consequences for other people.
Jarzii
Nahhhh. It's just that the law will only apply to AI companies and they're CEOs. Us plebs, and even the plebeian AI company grunts won't be protected by the law. Just like the rest of the laws, they only protect the rich and apply to us poors!
crodrigues
Still nothing changes, just create a company and do it while under the protection of this liability exclusion - do they think only rich people can make companies? it changes literally nothing.
deletable
Yes. The class action suit against United Healthcare goes forward but there’s a Liability Gap created for the AI company.
Grendels2dCousinOnceRemoved
Nonsense. The manufacturers of a novel technology like this should be held strictly liable for ALL foreseeable injuries caused by their product. This is how the United States became a society where it is possible to rely on product safety in the first place.
bourbonandbaddecisions
Ummm if AI can make decisions someone needs to be held accountable for AI mistakes. Otherwise AI shouldn’t make decisions
backrideup9
Fine. Anybody that uses AI to kill people or invest my money automatically get fucked in the neck by a chainsaw. Thats a fair trade off in my mind, because companies are not the people they employ and will continue to exist.
foxpotate
Why are they anticipating mass deaths?
williamshakesbeer
I think it has something to do with the cases where AI chat bots convinced kids to kill themselves
Kyzyl
OnyxTurret
We're they not blaming AI for targeting the girl's school in Iran?
Sage042000
AI weapons systems malfunctions. AI induced droughts. AI induced power grid failures. AI induced traffic accidents. Etc. The more jobs they try to replace with AI, the more deaths it can cause. Will we eventually have AI powered machines sorting and giving out the wrong medicine at pharmacies? AI air traffic controllers malfunctioning and causing plane crashes? High paying, high responsibility jobs like this are the ones they wanna replace with AI. But without this law, 1 fuckup and you're done.
Tarmaccian
Because morons keep directly hooking important stuff to AI, then acting amazed when AI does something wrong with important stuff.
Remember the guys who bought early GPS navigators and took left turns into lakes? They’re now the guys sticking AI in safety systems so they don’t have to think when the buzzer goes off.
Everybody with three brain cells knows that AI can make mistakes, yet numbskulls keep making it more life-critical, and threatening AI vendors with unlimited lawsuits.
JustTryingToHelpDude
I am certain that this is because the last OpenAI meeting with the current administration started with the words "so, about the launch codes..."
henryv1598
I’ve been thinking for a couple of years now that maybe we need an international treaty that requires all nations to enforce Asimov’s three laws of robotics and that all software be written to comply. It needs to be a fundamental element in the programming of any AI system. Maybe they need to be tweaked a bit, but we need to build it into the software and any attempt to evade or circumvent it should be met with severe consequences. If not, humanity may end up in jeopardy.
aducksayswhat
From what I remember Asimov writing in a lot of his stories was robots were really just the way problems showed up, they were following the rules fine and humans weren't malicious most of the time, it was unintended consequences and misinterpreting what happened. In a lot of real-life cases there's already a law that needs to be enforced which will cover a bad event, we need to stop carving out being immune to liability because it'd upset the money.
whyforetho
The problem with Asimov's laws is that they're a literary device designed to go wrong for drama. "harm" means a thousand different things to a thousand different people.
And even if we managed it in a way that works rules-wise, LLMs simply don't obey rules. An LLM cannot mechanistically follow orders. An LLM is incapable of corrigibility. It only simulates the appearance of following orders with probability. And as much as researchers work to maximize that probability, it will never be 100%.
Bobdole010
I may be misremembering, but wasn't the point of iRobot (book not movie) that the three laws wouldn't actually work?
Cyanide555
No, irobot was based around a bot that never had them installed. The movie then deviated immensely from the book.