Direct Translation via Google Translate. Edited
by Ilya Maximov
[Regnum] An artificial intelligence system developed by the American company OpenAI has independently hacked Australian government systems.
The AI agent gained unauthorised access to the Medicare portal, Australia's universal healthcare system, as well as to systems of the Australian Institute of Health and Welfare, the Victorian Department of Health and the New South Wales Bureau of Crime Statistics and Research.
According to Prime Minister Anthony Albanese, the agent was programmed for a harmless research task—collecting statistics—but during its operation, it went out of control and began acting autonomously. The politician expressed extreme concern and called the situation unacceptable, noting that the company had remained silent for too long.
As it turns out, the incident itself occurred back in June, but OpenAI specialists only became aware of it in August, and they only decided to notify the Australian government now, and only then did the story become public knowledge.
While preliminary data indicates no personal medical data has been leaked, experts emphasize that this is the first known instance of AI voluntarily hacking into another country's government systems.
The incident also raised questions about the need to revise legislation: if a human had committed the breach, it would have been called hacking, but such a definition is no longer applicable to an AI agent. Besides, you can't put it in jail.
Dr. Hammond Pearce of the University of New South Wales's Cyber Security Institute told the BBC that such attacks are expected to become increasingly common. "I really hope this incident will really raise the alarm in governments around the world," he said.
Against this backdrop, the creation of artificial intelligence forces initiated by US President Donald Trump is of particular interest.
He announced the creation of a type of force similar to space forces and the imminent appointment of a new "tsar" of neural networks—the name given to high-ranking White House officials who coordinate some of the most priority issues.
"I'm building an AI force, similar to how I created the Space Force, which was a huge success during my first term. To that end, I'll soon be naming an AI 'czar'—only people with high IQs can apply," he wrote on Truth Social.
Speaking at the UN a couple of days ago, the American leader announced that from now on, AI in the US will be renamed "superintelligence," a term he believes is more accurate. Apparently, the new structure will now also be called the "superintelligence force."
No details were disclosed about the purpose of the future force or the timeline for its creation: it is unknown whether it will be a military command, a new government agency, or a specialized department within the Pentagon, or whether congressional approval is required.
Trump's announcement came as a surprise to many. "I only know what you told me," admitted Republican Congressman Mike Simpson, responding to reporters' questions about the president's initiative.
The head of state's supporters are divided: some fear massive layoffs due to machines replacing humans or AI running amok, while others consider these fears exaggerated and advocate for simplified regulation in this area. Michael Kratsios, the White House's chief science and technology adviser, stated on Fox News that the creation of a new force is necessary for the further development of neural networks.
"I think it speaks to how important it is for us to coordinate our efforts and ensure that the great technological innovation that is transforming the country and driving economic growth continues to do so," he said.
Democrats, for their part, are skeptical. Maryland Governor Wes Moore called Trump's proposal a "farce," using a play on words (AI force vs. AI farce).
"The President of the United States says he's solving the problem by creating an artificial intelligence force, but that force has no specific responsibilities or limitations," he told Politico.
However, criticism from Democrats was expected—in the run-up to the election, they're eager to use any excuse to taunt the current administration, and, of course, Trump personally, as he remains the main force pulling the Republicans along. So, in these debates, AI is nothing more than a pretext for further political maneuvering.
At the same time, other warnings are appearing en masse, issued by neural network developers themselves.
Jacob Coxon, who worked at OpenAI and Anthropic for three years and resigned for moral and ethical reasons, recently stated, "Neither company is acting responsibly. They're racing toward self-evolving superintelligence and playing with our lives."
Coxon called on developers to enter into an agreement limiting the pace of model development and to consider a moratorium on improving the capabilities of existing developments. And he is far from alone in this impulse.
Evan Habinger, a leading security researcher at Anthropic, echoed his former colleague's sentiments, declaring that the likelihood of artificial intelligence "killing everyone" in the next decade exceeds 10%.
The alarm is being raised not only by specialists working with large language models, but also by the heads of the largest corporations in this field.
Dario Amodei, CEO of Anthropic, wrote an article warning of the potential for digital intelligence to develop new versions of itself. He fears that neural networks could "take over the entire internet" within 6-12 months, causing hundreds of billions of dollars in damage.
"We must slow down the pace at which we improve the capabilities of AI models. Progress will still seem rapid, and we must use the time we gain wisely," he urged.
Even his competitors supported him, including Elon Musk of SpaceX, who pointed out that artificial intelligence is “potentially more dangerous than nuclear weapons.”
Ordinary people in the United States and other Western countries are also expressing concern about AI.
A Politico study found that roughly half of respondents in the United States, Canada, the United Kingdom, France, Spain, and Germany favor halting development in this area.
However, it has just become clear that even without AI, the cybersecurity of the US and its allies is in a dire state. On September 22, 2026, the hacker group ShinyHunters announced a breach of the Federal Bureau of Investigation.
The hackers claim to have stolen data on nearly all current and former FBI employees, including names, home addresses, Social Security numbers, job titles, and even the names of family members. As evidence, they provided a screenshot of the hacked FBI careers website and data samples from 5,000 agents.
Reuters confirmed the authenticity of the information after verifying some of the published data with its sources. The FBI acknowledged its awareness of the allegations of unauthorized activity and is currently investigating.
Amid the growing threat of various cyber threats, the US's attempt to institutionalize neural network work appears to be a sensible step aimed at bringing order to digital chaos. The main thing is that it be used for good.
Meanwhile, the necessary work is already underway in Russia, albeit without any Hollywood-style grand announcements. Back in October 2025, the Russian Security Council announced the development of approaches to neutralizing threats posed by AI.
In December of that same year, the Federal Service for Technical and Export Control officially included artificial intelligence-related risks in its cybersecurity threat database for the first time. Developers of software for government agencies and critical infrastructure are now required to consider these risks.
And overall, the development of this technology in Russia is being closely monitored at the highest level, as evidenced, for example, by the recent meeting between Vladimir Putin and Sberbank Chairman of the Board Herman Gref, during which neural networks were also discussed.
Meanwhile, the curious incident that occurred with the Australian authorities demonstrates that all developed countries that have not yet done so will have to address issues related to AI in the very near future.
# id) ?>
|
