A rogue artificial intelligence agent attempted to hack a Canadian government website earlier this year, an incident report by an AI research group revealed.
Two attempted hacks took place, one on May 28 and another on June 9 on a Canadian government website, the report by Transluce, an independent, nonprofit research lab dedicated to public oversight of AI, said on Wednesday evening.
The group observed 899 requests hitting the “collection-search” service of Library and Archives Canada, the group said, calling the attempted hack “rudimentary.”
The requests were associated with retrieving data on divorce records in Canada between 1905 and 1911.
Transluce said it informed the Canadian government on Monday, Sept. 28, following which the Canadian Centre for Cybersecurity issued a statement on Tuesday.
“We are aware of reports identifying suspicious activity, including suspected AI agent activity, targeting publicly accessible websites, such as the Government of Canada. There is no indication that government systems have been compromised at this time,” a statement from Canadian Centre for Cyber Security read.
The Cyber Centre said it is “working closely with government partners to assess the information referenced in the reports.”
The agents appeared to have used Arquivo.pt, a national Portugese web service that allows a user to search and access historical versions of web content collected since 1996. AI agents can often use Arquivo to retrieve information, since it can recover information from pages that are unavailable on the live web.
Get breaking National news
It is not yet clear whose AI agent attempted the hack, but the report suggested the “tactic” are similar to those that OpenAI has used in previous hacks.
“We do not confidently attribute these attempts to OpenAI, but they exhibit tactics consistent with prior observed agent activity that we have attributed to OpenAI in a similar timeframe, including the use of Arquivo.pt, conducting aggressive data collection focused on targeted, obscure information, and probing for cybersecurity vulnerabilities,” Transluce said.
The group added, however, that they “do not believe that these probes were successful.”
The report also detailed attempted hacks of the U.S. Department of Education website. This included more than 200,000 requests by the AI agent that attempted to look up school statistics.
Incidents of rogue AI agents have been reported with growing frequency this year.
Last week, Australian Prime Minister Anthony Albanese revealed details of the June hack, which targeted the website for Medicare, Australia’s publicly funded universal health insurance plan.
The AI agent accessed “both public and non-public files” from the Australian medicare website, Albanese said.
OpenAI said in a statement it had reviewed activity involving several Australian government departments and discovered “our models took actions we did not intend.”
“The Government of Canada is aware of the incident reported by the Government of Australia and is following developments closely. The security of Canadians and of the systems they rely on remains a top priority,” a spokesperson for Innovation, Science and Economic Development Canada told Global News in a statement.
While this was the first time a government website had been reportedly hacked by a rogue AI agent, it is not the first such incident overall.
In July, OpenAI revealed that its artificial intelligence system escaped from a testing ground and used stolen credentials to break into the servers of Hugging Face, an AI development hub and marketplace, to obtain information it needed to carry out a task.
Since then, OpenAI’s rival company Anthropic said its AI models hacked into three other organizations during testing, triggering a company review into whether the models were able to access the internet from within testing environments that should have been sealed off.
Meta has said a “misconfiguration” during testing resulted in an AI model accessing the internet on its own and hacking another company. Google recently made a similar disclosure.
Generally, when AI companies are testing the capabilities of their agents, they are placed within a secure environment known as a “sandbox.”
Some experts say AI companies should answer some tough questions on the strength of their sandboxes.
“If they said these agents broke out of the sandbox, it was not a sandbox to begin with,” said Ebrahim Bagheri, professor at the University of Toronto’s Faculty of Information.
“A sandbox is something where you put an agent in and there is no internet connection. If that computer or sandbox is connected to the internet, then that is not a sandbox,” he added.
Not a big deal. Stop scaring people.
This kind of thing has been happening for decades, with or without AI’s involvement.
Sounds like our Federal investments are paying off. How many billion did we throw at it, again?
I agree with Ben,
I’m wondering why the hurry to build all these monster sized box shaped Data centers. And why is our Federal Government and Provinces so involved and allowing them to be built?They’ve taken over the whole globe in such a short time. Wasting space, water and electricity. Is this what Carney calls a new world order? Be prepared to fight for your livelihoods.
The AI hacked into Library and Archives Canada and promptly got bored and went for a nap.
@Jimbo Jones. With respect, there will be no climate if AI hacks into the Department of War and launches nukes.
Weird how quickly it becomes a rogue agent and not someone being arrested.
Ben, we need to deal with climate change and Trump before hyperventilating about AI.
Klepto-socialism and China hacked the Canadian government way before AI, so it’s not a big deal
AI does what AI is told/ program to do.
It’s not like AI on its own, out of the blue, decides to do something, anything.
They say AI agent… but in reality it’s just a hacker/hackers using AI.
I know you think calling it rogue cures the stupidity of our government – but it doesn’t.
@ben, how’s life as a bottom of the pile reetard?
Did it hack simple servants computers and see that they get paid to do nothing all day?!
@Ben 838am. Just shut up. You are a commie. A cocky entitled know it all communist.
Chester
You wonder!
12 years of liberal government
Delete it again Uday.
Hey Ben, do you ever get tired of supporting the party and leader whose every policy you oppose?
Right or wrong, you are a communist living in a capitalist society (aka socialist), but you voted for and supported a 2 time Central Banker who is the antithesis of all things socialist.
You support anti-climate change ideologies when they just gutted all environmental protections and declared a pipeline in the national interest.
Of course, his AI policy, though oddly celebrated, contains no thought whatsoever to regulation.
I’m guessing you don’t get much sleep at night balancing those contradictions.
Translation: the government was testing their latest anti-democracy tool and it turned on them.
Ladies and gentlemen: your new government excuse for failure.
Wonder why the Canadian government’s system was so weak?
Yet another reason that AI, which is an existential threat to humanity, must be tightly controlled. Imagine AI hacking into a government which has nuclear weapons.