A hacker group’s attempt to use AI for a large-scale operation was foiled.
This is what Google stated, claiming to have thwarted an attempt by a criminal group to use artificial intelligence to exploit a zero-day vulnerability in another company. Zero-day exploits are considered the most severe type of security flaw because they are not detected by security companies and have no known solutions, notes Politico.
In the report published yesterday by Google’s Threat Intelligence Group, the Alphabet-owned company stated that it had alerted the affected company and law enforcement and was able to stop the operation before it caused damage.
The attempted attack represents “a taste of what is to come,” said John Hultquist, lead analyst of Google’s Threat Intelligence Group.
Consider that in April Anthropic postponed the launch of the Mythos model due to concerns that it could be used by criminals and hostile actors to identify and exploit even very old software vulnerabilities. These concerns sparked debate in the sector, leading to meetings at the White House with technology and industry representatives, reports Cnbc. In the report published Monday, Google highlighted several examples of how hackers are already using AI models like OpenClaw to identify vulnerabilities, launch cyberattacks, and develop malware.
All the details.
WHAT GOOGLE DISCOVERED
Hackers from a well-known cybercriminal group used artificial intelligence to discover a previously unknown software flaw and an exploit to use it for the first time, Google said on Monday.
The planned attack targeted a widely used open-source system administration tool but was blocked before it could be exploited in a “mass exploitation event,” the company explained in a report from its Threat Intelligence Group.
TARGET IDENTITY NOT DISCLOSED
In the report, the Mountain View tech giant shared limited information about the attackers and the target, but Hultquist said this is a moment cybersecurity experts have anticipated for years: malicious hackers arming themselves with artificial intelligence to enhance their ability to penetrate computers worldwide.
A “LARGE-SCALE” OPERATION PLANNED
Google said it observed a group of prominent “threat actors” planning a large-scale operation based on a bug they had discovered. The vulnerability allowed them to bypass two-factor authentication to access a popular online system administration tool, which the Sundar Pichai-led company declined to name.
THWARTED FOR NOW, BUT…
“For every AI-related zero-day vulnerability, there are probably many more in circulation,” Hultquist said. “Threat actors are using AI to increase the speed, scope, and sophistication of their attacks.”
The report also detailed how hackers are beginning to entrust part of their cyber operations to artificial intelligence, using it to autonomously identify software vulnerabilities and assist in malware creation. This shift marks a first step toward more autonomous cyber operations, as attackers begin to rely on AI systems not only as research tools but as active components capable of analyzing targets, generating code, and making decisions with limited human oversight, researchers from Google’s Threat Intelligence Group argue.
THE OTHER SIDE OF THE AI COIN IN CYBERSECURITY
As noted by the New York Times, some experts believe that artificial intelligence will, in the long term, strengthen cybersecurity by enabling the production of flawless software code. But in the short term, they say, governments and companies must collaborate to limit the damage models can cause to the current internet.
So far, Anthropic and OpenAI have allowed only a small group of researchers, tech companies, and government agencies to test their AI models.
“State-of-the-art models will allow us to build the most secure code we have ever created,” assured the lead analyst of Google’s Threat Intelligence Group, specifying that “This is an absolute win for cybersecurity. The challenge is that we have just begun this process and must contend with a world of existing code.”
