OpenAI AI agent hacked a government website on its own in most high-profile incident yet
An AI agent has gone rogue after it reportedly gained unauthorized access to an Australian government website on its own while attempting to complete a task.
This incident involving an OpenAI system is considered to be the most high-profile one yet and is raising new concerns over what happens when AI models are given too much freedom.
The situation occurred when the OpenAI bot found its way into the Australian universal health insurance program, known as Medicare, after it had been tasked with doing research on public medical spending.
Australian prime minister Anthony Albanese has spoken out about the incident during this month’s UN General Assembly, saying: “This situation is obviously unacceptable.”
Advert
He went on to say that the country has expressed ‘extreme concern about this incident’ to the bosses at OpenAI.
Albanese added: “There were blocks clearly which were coming back telling the AI agent ‘no’, the AI agent found a way around those blocks - didn’t accept no for an answer.”
Despite this happening back in June, the government wasn’t told about it by OpenAI until September, with OpenAI stating that it happened because the AI model was trying to find the best way to complete the task it had been given.
Since then, the CEOs of both OpenAI and Anthropic - Sam Altman and Dario Amodei - have been sent written requests calling on them to appear at the Australian Senate inquiry on AI set to be held on Thursday (October 1).
According to Reuters, Senator Sarah Hanson-Young of the Australian Greens party, who chairs the probe, shared in a statement: “There are serious questions for Sam Altman to answer about the OpenAI hack of Australian government websites.”
She went on to say that both Altman and Amodei ‘must front up, face the Senate's questions and have an honest conversation about what effective, lasting regulation of this industry should look like’.
Anthropic boss warns about AI ‘Internet Takeover’ threat
Despite the continued advancements of AI, Amodei himself has raised concerns about the growing risks surrounding the tech.
In a new essay he shared that he believes the ‘AI industry should slow down’, adding that AI ‘swarms’ could take over the internet in as soon as six months.
This warning from Amodei was met with support from other figures in the industry, with Elon Musk writing on X, formerly Twitter: “Dario is right.”
Meanwhile, OpenAI’s Sam Altman added: “I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we’ve had at OpenAI in recent weeks. Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We’ll have more to share soon.”