So apparently AI is “hacking” websites. Translation: An AI was prompted to red test a website and got into it.

Australian authorities have raised the alarm after OpenAI-powered models hacked into a government health data system in June, slipping past its digital defences and accessing files without authorisation. This is the first publicly known case of artificial intelligence (AI) “agents” – AI-powered software systems that can carry out tasks autonomously – breaking into a government website, and the latest of several AI breaches of external systems.

STOP! full stop here. This is not the first nor the last, AI has been used to pressure test websites for more time than we can count here and if you load a AI with every way that they can leverage a website the premise of hacking is a lot less hacking and more site/user error. People need to stop acting like AI is the best hacker on the planet because it is not. It is the best average leveraging agent meaning it will use and abuse all of the common ways we screw up as humans.

This article is just another continuation of the AI hysteria that AI devs have been trying to drop on the populace lately.

The disclosure comes as top AI firms warn of the risk of humans losing control of AI, calling for its development to slow to a pace that allows it to be safely regulated. Global powers must cooperate to ensure this, they have said. A research scientist at AI firm Anthropic, Evan Hubinger, went so far as to say he believes there is a greater than 10 percent chance AI could “kill all humans” within a decade.

Look, as much as some of these armchair scientist at Anthropic keep saying this it gives them free press. I have a solution and it’s cheap. for $28 you can go down to the home depot and get a 600 Volt Industrial Grade Three Pole Three Phase AC Toggle Switch, You save humanity just by using 0.01 to 0.05 Joules of mechanical energy. That is all it takes to save humanity from killer AI.

Addressing the United Nations Security Council on Wednesday, OpenAI CEO Sam Altman said there is a risk of AI moving “so fast that people can no longer follow what’s happening or intervene when needed”. “This would obviously be terrible,” he said. “And we should not train models that we cannot make an extremely strong case that we will be able to keep under human control.”

This shit needs to stop, Mr Altman and Mr Suleyman need to fuck the hell off. They are not working on a Manhattan project, they are refining AI that is years old and they know exactly how much it can do. The only change has been the amount of material the LLM has at its disposal. Instead these jackasses are out saying judgement day is coming, why, because it’s marketing. You want to shutdown an AI EMP the SSD and its gone or go the cheaper route and and go to home depot.

When OpenAI accessed the government portal while conducting research on public medical spending, Albanese said the AI agent circumvented “blocks” that should have prevented it from breaking into the portal.

This was not a hack, The AI knew the structure of the dashboards and likely the aussies never secured a dashboard that they thought it was not in public view. It likely tested /admin/, hit a wall, and moved to /sadmin/ where someone forgot to enforce authentication… Most of the medical CMS platforms are built in secret and are not tested with a public face and likely this AI has seen every stupid mistake that every base level user makes and used that. So this hack was not magic it found a tool.

Deputy Prime Minister Richard Marles said the information the OpenAI agent accessed was “not particularly sensitive” and was later publicly released. Still, Albanese called the situation “obviously unacceptable” and said Australia had relayed its “extreme concern” to OpenAI, which had failed to notify the government of the breach until September 10.

When you give someone or something the average way the world works around security it is bound to find a way in. You want your way around a master lock . go to youtube the lockpicking lawyers hands will happily tell you how. You want a way to get a file back, Again youtube. Technically humans have the same leverage to learn ways around things just with Youtube alone. So before calling AI a superweapon or Master hacker look at youtube.

In a statement, OpenAI said it had “identified activity involving several Australian government websites and services as our models attempted to look up answers” and “took actions we did not intend”.

This is a case of the answers are normally “HERE” and the AI took the correct route to the control panels and it got information that was PUBLIC and has been misconstrued as a hack.

Yes. The Australia data breach is the latest of several instances in which AI agents belonging to OpenAI, Google or Anthropic have accessed external systems without authorisation.

OK… Did someone not proofread? Earlier it said ” This is the first publicly known case of artificial intelligence (AI) “agents” – AI-powered software systems that can carry out tasks autonomously – breaking into a government website, and the latest of several AI breaches of external systems.” UGH … Really … No .. I’m not even going to comment on this . My comment is “comment” for this someone call captain obvious.

Ed note: most times AI’s probe natural addresses and go to /admin/ /dashboard/ etc., which is not hacking but logical probing to find landing pages for various things. The web is only as secure as the person using it.

Attributions From:
Aljazeera – How an OpenAI ‘agent’ hacked Australia’s Medicare and what that means