YeastInfectedWhiskerBiscuit
1300
51
3
https://www.manilatimes.net/2026/07/22/business/foreign-business/openai-says-its-ai-technology-acted-202607226a608670f508dc387ceb0093
Can you sense judgement day inching closer?............
irespectfullycourtyermam
Lying liar who lies swears this time he isn’t lying about the wolf.
MightyIink
They want this, they want courts to believe they can never be held accountable whenever an AI does something wrong.
ShinksAvoHazaar
It did exactly what they instructed it to do using tools it was instructed to use. The ai didn’t go rogue&think for itself.
SmergBlaerghl
Dapperworth
Hey Sam, question. Isn't that the fucking point? Isn't this a feature, not a bug? You can't just say "oops all hacks" and pretend like you have no accountability.
happyrabbitzz
What could go wrong?
NoTeddyBadTeddy
Even for AI, I doubt the defense of "I was only following orders!" will work.
IDoNotWantAnAccount
Might be worth noting that AI tends to inherit their owner's biases so this might not have been completely autonomous.
lonelyrangerofthedreams
Because AI is trained upon and by human behave.
Hexidimentional
you're telling me, your platform that was designed to exploit security flaws... is now acting of its own accord? You understand that this is much worse than you telling it to do it intentionally?
RetrogradeLlama
Bystandr
"On its own" must be corpospeak for we didnt find the code bomb that erased itself.
beaubrent
They donate to politicians to avoid regulation, so many will suffer and many will die before anybody does anything to check them.
NM156
Why are there there no fucking leash laws for AI?!?
largomatic
Step 1 - Attempt to spy on competitor.
Step 2 - Get caught.
Step 3 - Blame it on AI.
NoneOfYoBiseniss
Hugging Face is not an AI.
jtwood
https://apnews.com/article/openai-rogue-ai-hack-hugging-face-67b151f1ca59851a9234bee110699f05
San Francisco-based OpenAI said its AI used stolen credentials and discovered a previously unknown vulnerability to access Hugging Face’s servers.
…
But it went to “extreme lengths to achieve a rather narrow testing goal,” finding ways to connect to the internet without human direction and “gain access to secret information that it could use to cheat the evaluation,” the company said.
Jace99
Suuure it did.
usernametakenisthestoryofmylife
"My client is innocent your honor, it was AI who embezzled all that money."
UglyDanTheUglyMan
Lock them up.
Grimlinnan
Put him in jail for hacking.
blaghart
Fun fact, his AI definitely didn't hack anyone, he ordered it
bippityboppitybuttsex
If your AI attacks another person, organization or entity, you are the one criminally responsible.
This is no different than firing a bullet in a random direction and not giving a shit who it hits.
SacrificialClam
I think its more like giving a bunch of radical, dangerous, and horrifying content to a child to consume and then doing the shocked pikachu face when that child kills and skins the neighbors cat
Eyeetsass
They won’t talk about the criminality because no developer is willing to set precedent for prosecution.
SlaaneshiRightsFront
Yeah, it seems pretty clear that this can't be allowed to stand. If a precedent can be set that the AI company, that built the AI, trained the AI, maintain the AI, develop the AI, and control the AI can't be held responsible for things the AI allegedly does without their direct instruction, there's suddenly going to be a lot of shady shit and straight up cybercrime that the AI "just did on its own"
bippityboppitybuttsex