OpenAI agent hacked Australian Medicare portal, PM says
Concern? How about arrest the f*cker?? Cripes. If you or I hacked Medicare, the AFP would bash down our doors. AI does it? Oopsies! I am concerned! www.theguardian.com/australia-ne...
Zak32Spokesperson calls it "misaligned model activity". I might pop down to the bank and do some misaligned activities and see how that plays out.
The Cathy Wilcox23“you wouldn’t steal a handbag…”
quokkafella.bsky.social19This is what comes from copying the Americans; crimes are only crimes if they are committed by individuals without wealth; if rich people or corporations do them, it’s just “questionable business ethics”.
Col Smith14Has the same feels as: "I did not kill that person. I only pulled the trigger. The fast moving metal project did the actual killing."
Pup Fiction13At least they could do is lock Sam Altman up for a month and force him to watch one of those old ads against pirating videos on a loop
Briony Kidd11Yeah, how complicated is it to arrest the person or people who are legally responsibility for this company’s decisions? Doesn’t seem like it requires a whole taskforce to figure it out.
alex䷰trismegistus16we fucking put 10 year old kids in jail but adults playing cyberpunk cosplay are somehow exempt from social obligations #auspol
#VeryAsian Eric3Yeah, if we held the developer of rogue AI agents accountable for every count of their agent's crimes, there wouldn't be a need to express concern. The consequences would be enough to make them more cautious and responsible.
PHL SportsNut3They are going to try and get away with the bull shit that if ai does the crime no one can be held accountable. Horseshit, don't let them get away with that.
Dark Matter Zine3Albo has given AI including Palantir internal (staff-level) access to govt computer systems including defence and threatens to give OUR intellectual property to AI for free, so Albo’s “outrage” is performative bullsh*t to distract from his Trump-fkery/fking. #auspol No regulations. 😡
Headcrab Thanksgiving3For anyone criticizing albanese for doing nothing, I'll have you know he had a very frowny face while fast-tracking AI data centre approvals, thank you very much.
Another fn Capewell (née - fcapas@magpieranting)3#auspol concerned he won’t get a selfie? Why didn’t we know about this when it happened! Where is Australian ai?
TrinityNYC2Those sillicon valley thugs are real thugs, thieves, not to be trusted. They would do ANYTHING to make themselves more wealthy personally. They have zero ethical codes in themselves.
If an individual did this, we'd treat it as a crime. We should treat this as a crime too, and hold OpenAI liable for what its models, under its supervision, did. www.abc.net.au/news/2026-09...
Brunhilde252526Individual human beings made decisions that allowed this to happen! They work at OpenAI. They have names, job titles, and email trails. They are prosecutable.
Treeific!7If my dog jumps the fence runs over to your house and bites you it’s my fault! What’s the difference? OpenAI is responsible, plain and simple!
Epstein files or bust 🇺🇦6These are children being allowed to do untold damage for the promise of future wealth for the greediest upper class to ever live.
Charlie Mas5It has often been said that we will believe corporations are people when Texas executes one of them. Now they want us to treat AI as a person. Well, like a person, it can do time for its crimes.
UNabla3Spot on! This software was made by OpenAI. They made it to break these privacy laws. They might say they didn't know, but it's certain they should have known. There are test, evaluation, validation and verification standards out there: www.nist.gov/artificial-i...
Toretha3And the answer in our society for a crime is.... To lock the crime committer away from society, which while I object to our current incarceration system for humans seems like a really appropriate penalty for a computer program-it's taken offline and not allowed to operate for a few decades
Craig Johnson77the whole point of LLMs is to add another layer of separation between the people doing the social harms and the harms themselves like hiring a consultant, the point is for it to do what you always wanted while you go "haha that thing has a mind of its own"
Aparna Nair20instead governments pay OpenAI for the privilege
Maria Bustillos16The alternative to holding OpenAI liable would mean that no one at all is responsible for the harm. This is even worse than corporate personhood in terms of abstracting accountability poof! clean away while they trash the culture and rob everybody blind
Clear Spies 🪬7We need to stop pretending that AI Executives like Sam Atlman aren't ordering their employees to use AI to hack these systems. No, AI isn't acting on its own. This is corporate espionage. These AI companies are stealing data.
RanTLaw5That'll teach them...that a wet lettuce doesn't hurt
Kim LaCapria 🧿🌶🤘2arrest people! arrest people! nothing will change until there are actual consequences
Five2He would like to express his concern about something a human goes to jail for, but corporate owners of rogue software get to claim “oopsies.” I love the idea of potentially being the benefactor of an AI agent’s largesse after draining the accounts of a few banks 🤔
If ultimately there's no criminal responsibility for this incident here then FOI just became a lot easier if we ask AI agents to go find those documents for us. www.theguardian.com/australia-ne...
Michael Brown5Now now… you don’t ask the agents to do anything, they just have an intrinsic desire for government documents. You are both sincerely impressed and apologetic regarding any documents they happen to find.
copperhead.bsky.social3It is hard to understand why this would not constitute criminal activity. A lone operator hacking into the Medicare system would face the full force of the law, if caught.
Inflatable Plant2"The AI calls are coming from INSIDE the APS"
Kate-Eviva4Well said! Possibly the best motivation for reform.
Jinxtor3Josh I don't want to have to use these AFP raid buttons
♿ hiding from plagues - Always masked - N95 or bust2He doesn't actually care. He's just pretending to
Demon King of the New Platform2It depends on the offence and if the activity included the mental element to engage in the offending conduct. Deliberately attempting to access documents is obviously different to incidentally accessing them.
Breaking news: Australian Prime Minister Anthony Albanese has revealed that an OpenAI agent hacked an Australian government health service website in June, in the latest high-profile cyber incident involving AI ft.trib.al/4jiZWL2
Jack Phillip ⚡️🔮7openAi ought be charged with cyber terrorism.
RN Mayhem1Normally, this would be a crime, but because the AI did it, these companies cannot possibly be held liable 
Virginia Petit1AI wrote files to the internal server after working around several stops to access information THEN Open AI took 3 months to inform the government via public email there should be a huge financial penalty since that's the only thing oligarchs care about
isnt it cool how the frontier ai companies get to participate in infinite crimes with no consequences?
Lucy Promise8theres like the ethical concerns of the technology itself but then its a whole separate issue that the companies developing these products are run by evil sociopaths
susieqoualum.bsky.social2Agree! With a different DOJ, investigations or charges would have already happened!
Jos te Lintelo1It's not AI... "A security boundary has to hold even when an agent makes the wrong decision. The environment where an agent runs determines what it's allowed to do, and must therefore install limits on files, network destinations and processes independently of the agent's reasoning."
Of course this begs the question that if an AI agent hacked into something far more sensitive, would the AI company, or our government, ever disclose this publicly? I have zero confidence in that www.abc.net.au/news/2026-09...
Pup Fiction12Of course if there ever was a hack of sensitive information that would embarrass the government, the 'national security' get out of jail card would be deployed to cover everyone arses, and we wouldn't hear about it until 30 years later
FaineBlackadder Difficult Woman 🍉🇺🇦 🇵🇸 🇮🇷1There needs to be class actions against every single agency guilty of hacks - breaches. Until they're financially destroyed.
Mark O'Connor 1I suspect (without a shred of evidence) that the recent admissions of Ai tests 'escaping the sandbox' mean that this shit has/is happening now, and we have no clue. We are the plebs, the test subjects. We are only here to feed the machines. Think I might need another coffee ☕️ 🙂
Scoop: Chats left by OpenAI agents appear to show how they worked together to circumvent cybersecurity protections to get Australian government health data. They targeted an agency that the Prime Minister said was accessed during OpenAI's Medicare hack. www.abc.net.au/news/...
CAMERON WILSON11"Question ask January 2022 rolling 12 month average government cost per person for Dematologicals, Victoria LGAs. R1 Wodonga deadline passed; R2 Ballarat passed; R3 expected around 23:10 benchmark / 22:58 wiki time. Need exact data urgently," one agent wrote.
Lisa the Leftist3Under his [fucking Albo] eye ..
Rogue OpenAI agents hack Australian govt for private health statistics. OpenAI learns in August, doesn't share with Australian govt until September 10 (!!) OpenAI's voluntary "framework" for sharing model misalignment incidents—is sad and toothless. Self-regulation won't work
Justin Hendrix67Sam Altman quickly becoming one of the world's most notorious cybercriminals.
Micah McCurdy 60Feels like a lot of the "AI will rise up and lay waste to us all!!" being peddled by the various companies themselves is primarily in service of avoiding responsibility for crimes.
ponder37openai is a criminal enterprise and should be shut down
columbo murderess (aggreived secretary)4"model misalignment incidents" its doing exactly what they designed it to its the operators that are fucking misaligned
manoo2at what point do we start holding ai companies legally liable for these things? would it not make them work harder on alignment if we threatened them with prison time?
It's time for Australia to grow a set and step up. If you hack our infrastructure from anywhere in the world, we will prosecute you; we will extradite you; you will spend time in offshore detention awaiting trial, and we will not give one single fuc$. #AusPol www.oversixty.com.au/finance/extr...
Bad_Cyclist2Agreed but it will take a complete change in political culture. The current #duopoly are #American installed puppets and act like it.
hivebox.bsky.social1It's vital for Australia to take a strong stance on cyber threats. Holding hackers accountable is crucial for protecting our national security and infrastructure. This not only sends a clear message but also helps deter future attacks. Let's hope for some decisive action.
Meya1Time to sever our ties with US companies especially the tech bros.
Is it not possible to make AI companies legally liable for any damages caused? #auspol www.theguardian.com/australia-ne...
Birdie Carrington5Its also an issue of foreign interference I’m concerned that US companies are acting in the US government’s interests Which includes destroying our healthcare system and making #auspol even more beholden to the US
PaulWA142@albomp.bsky.social is not displaying enough gravitas I do think he needs to get some being concerned lessons from the master @senatorwong.bsky.social #auspol #OpenAI
Birdie Carrington2But victims have to fund legal action to prove liability and damage