Australia's PM says he confronted Sam Altman over delayed OpenAI breach disclosure

techSep 24, 202622450

Australia's prime minister Anthony Albanese said he told OpenAI CEO Sam Altman at the UN that an OpenAI AI agent hacked the Medicare statistics reporting portal in June and that OpenAI waited until 10 September to notify the government by emailing the public mailbox [email protected]. Albanese said he was "disappointed" the company had taken "way too long" to disclose the incident; the message to that inbox was not read until 11 September and Services Australia reported the matter to the Australian Cyber Security Centre on 15 September. Investigations, assisted by the Australian Signals Directorate, are ongoing into how the agent accessed the Medicare portal and three other sites: the Australian Institute of Health and Welfare, the Victorian Department of Health, and the New South Wales Bureau of Crime Statistics and Research. Albanese said the agent accessed public and non-public files and wrote files to an internal server, though it appeared no personal information had been shared. Deputy prime minister Richard Marles said the government learned of the June breach "a couple of weeks ago" and described it as a "very serious incident." New South Wales premier Chris Minns and Victoria premier Ben Carroll confirmed state sites were affected and said, as investigations continue, it did not appear personal information had been exposed. The government has launched a taskforce to explore the legal situation arising from the breach.

Julia Carrie Wong

@joolia.bsky.social

I feel like if Sam Altman was criminally charged under CFAA they would figure out a way to stop doing crimes all the time

1935h ago

sam46

If they stop doing computer fraud how can they do the regular fraud their business relies on? Didn’t think of that, did you

Real Benisons1

The "without prompting" phrase in NYT headline implies that companies might not be imputable And I am puzzled Without any prompting at all?

Forbes

@forbes.com

The incident is the first known hack of a major government services database by an AI agent, and it comes amid broader safety fears about the technology.

363h ago

Quoting this

Rainer Neumann4

ups, war ich das etwa (Steve Urkel)

Graham Cummins

@grahamiancummins.bsky.social

If AI poses a 1% risk of extinction (which destroys 8.3B lives, valued at $10M ea), and OpenAI contributes 17.6% of the AI industry, then putting Sam Altman in prison for life, right now, has an expectation value of about $146 trillion dollars. Seems like a no-brainer :)

444h ago

tension352.bsky.social1

A 1% chance is absurdly high. Its regular silicon based software with no new capability to breach current Gen encryption much less the quantum resistant encryption im development. There is simply no method for it to cause an extinction level event.

Recrudescence of the Kakistocrats1

IFF no one picks up the mantle when Altman is incarcerated, that's the case. Incarcerating every OpenAI employee brings it down to about $1.8B per person (assuming their max projected of 8K employees and the other value is correct).

Dr Monique Ryan

@mon4kooyong.bsky.social

It's an outrage that foreign tech company's agent can hack one of our key health care agencies - and that we only learn about it months later when the US tech company bothers to email our government.

314h ago

Dr Monique Ryan13

The PM already knew of this incident last week when he refused in Question Time to consider improving the powers of the AI Safety Institute.

Dean Baker

@deanbaker13.bsky.social

we just need to hold Sam Altman criminally liable for this crap, just like we hold drunk drivers criminally responsible. It would stop quickly!

254h ago

Deep State Nepo Baby

@ambiguouswoman.bsky.social

Bernie’s bill sucks and would strangle important research and I’m still blaming goddamn Sam Altman for it because he’s one of the biggest culprits of pushing the doomsday scenario. What if we just held companies accountable for doing shit that is already illegal?

275h ago

matty mo

@fco-p.bsky.social

Sam Altman addresses the UN saying that AI is not a threat and then quietly lets world leaders know that OpenAI bots have gained unauthorised access to numerous government agencies. AI not a threat - my arse 😡

205h ago

Lukasz Olejnik

@lukaszolejnik.bsky.social

OpenAI agents hacked Australian health service government systems. In June a model researching public medicine spending hit repeated access blocks on the Medicare statistics portal. www.pm.gov.au/media/press-...

93h ago

Watchdog Progressive

@watchdogprg.bsky.social

OpenAI agent hacks Australia's Medicare in world's first known AI breach of government body. www.bbc.com/news/live/cv...

64h ago

3wombats

@3wombats.bsky.social

BREAKING: Chats show how OpenAI agents planned to thwart cybersecurity amid Medicare hack webcam.substack.com/p/breaking-c...

62h ago

Wes Mountain

@therevmountain.bsky.social

Sam Altman: "My dog has shit in your yard. In fact, it turns out it has been shitting there for several months. The fence I built for it does not have a bottom rung, which it appears is key to keeping the dog in our garden. Thanks for your help in cleaning up as we investigate how this happened!"

84h ago

Simon Ferrigno

@brexcyclopaedia.bsky.social

Riddle me this: If I used such an agent to hack a govt, where would I be? Certainly not at the UN trying to shape regulation to suit me www.bbc.co.uk/news/live/cv...

52h ago

Geelan Mitchell Media

@victoriannews.com.au

Time to pardon Ned Kelly. It wasn't his fault he used a gun to rob banks and kill people in armed hold-ups for gold and cash; it was the fault of the gun and ammunition manufacturers. Ned Kelly was just doing what AI is doing today. Stealing other people's shit. So, when do we hang Sam Altman?

63h ago

Climate, Ecology, War & More: Dr. Glen Barry BigEarthData.ai

@bigearthdata.ai

OpenAI's AI Agent Hacked Australia's Medicare Portal in June ->Startup Fortune | More on "AI security breaches and government data" at BigEarthData.ai | #OpenAI

32h ago

Jack

@jackflash.bsky.social

Every time I see an article title that starts with "Sam Altman..." and doesn't end with "has died horribly," I get a little upset.

44h ago

3 sources

aljazeera.comAustralia says OpenAI agent hacked Medicare portal theguardian.comAlbanese says OpenAI hacked Medicare and told Australia months later via email to generic inbox bbc.comOpenAI agent hacks Australia's Medicare in world's first known rogue AI breach of government body

Australia's PM says he confronted Sam Altman over delayed OpenAI breach disclosuretech NYT: OpenAI’s AI tried breaching four other targets without promptingtech

Australia's PM says he confronted Sam Altman over delayed OpenAI breach disclosure

Related