by Steven Richards
New AI-powered tools and autonomous hacking frameworks are purportedly accelerating large-scale cyber attacks, prompting urgent warnings from the FBI and tech leaders. Google’s Threat Intelligence Group calls it "a critical inflection point.”
Amid service outages and conspicuous failures in digital systems across the United States, the FBI is warning about new AI-powered tools being used by “state-aligned hackers” to conduct large-scale cyberattacks on vulnerable digital infrastructure.
The official FBI warning follows a string of digital infrastructure failures, including only the third nationwide grounding of flights in U.S. history. Also, system crashes have disrupted operations at United Airlines, Microsoft and major banks, and hacks of major municipal water systems have occurred.
At the same time, Google is now warning that new Artificial Intelligence (AI) tools are giving hackers the capability to deploy autonomous agents to find and exploit major vulnerabilities, engage in theft, and conduct information operations.
AI-assisted hacking is becoming more autonomous and faster, FBI says
“[T]he adversarial AI landscape has reached a critical inflection point,” Google’s Threat Intelligence Group concluded in a report published Monday.
The FBI and private sector investigators are both warning that the latest AI intrusions and observed behavior show that AI-assisted hacking is becoming more autonomous and faster as AI-powered tools are able to take on more complex tasks. These tools are being exploited by state actors and criminals alike, according to the FBI.
The FBI issued a FLASH advisory last week warning that “state-aligned hackers” were using an AI-powered tool to conduct large-scale cyberattacks on vulnerable digital infrastructure, Just the News reported on Tuesday.
Specifically, the bureau said both cybercriminals and state actors are using ARTEX, “a superintelligence-powered penetration testing framework,” in order to automate attacks against businesses and organizations. Known as ARTEX, the open-source model uses AI models to conduct thousands of entry attempts with minimal prior expertise and without manpower limitations.
Method in use since at least August
Trade journal IndustrialCyber reported that the FBI said malicious cyber actors have used the framework since at least August to target organizations across financial services, government, technology, defense, semiconductors, energy, healthcare, manufacturing, telecommunications and legal sectors. Some intrusions progressed from initial targeting to privileged access in less than three hours, the publication reported.
“ARTEX significantly reduces the time required to chain attack techniques, with some intrusions progressing from initial targeting to privileged access in under three hours,” the FBI’s Dallas office explained in a Tuesday post to social media. “The framework enables threat actors with minimal technical expertise to conduct complex, multi-target operations simultaneously, fundamentally accelerating attack velocity and expanding operational scale,” the FBI added.
The bureau found that Chinese models are the most prevalent in these kinds of attacks, especially the open-source DeepSeek Flash family models, in part because of their “lack of observable guardrails.”
Bassem Youseff, former unit chief of the FBI’s counterterrorism division, told Just the News that “the bottom line is that we've known for years that China is very adept at hacking, and the fact that now they are using AI just increases […] our need for us to be more ready than we ever have been.”
“I'm thinking […] back to the days of the idea of denial-of-service attacks, which was the big thing in those days. Think of that on steroids, where basically, AI expedites and expands this and speeds up the process of actually conducting these attacks, or for the NSA to be down for a period of time, and of course, all that we've seen for the last few days, you can't disregard the fact that China—now, of course, we don't know that—but China and Russia, obviously, even with Iran using their technology, using Chinese technology, would be a culprit for sure,” Youseff told the Just the News, No Noise TV show on Tuesday.
Although experts say China and Russia have the capacity to launch these kinds of attacks, as of press time there have been no arrests or suspects named, either foreign or domestic.
Intrusions are now “occurring at machine speed,” Google says
Google also warned about how “threat actors” are using ARTEX and other similar platforms to conduct penetration testing to detect vulnerabilities in systems.
The company’s threat intelligence team said this reality means intrusions are now “occurring at machine speed” as threat actors learn how to automate tasks and phases during intrusions. Google’s Threat Intelligence Group called this development a “paradigm shift.”
Additionally, hackers are now using AI models “to continuously scan software updates, spot newly fixed bugs, and build attacks before teams can apply the fix—all without human involvement,” the company said in a post to X.
In recent months, AI company leaders have publicly sounded the alarm about the rapid development of AI capabilities and the threats that could stem from new capabilities. Open AI CEO Sam Altman, for example, agreed that a “world-shaking cyberattack” was “totally possible” within the next year.
“I suspect in the next year, we will see significant threats we have to mitigate from cyber,” Altman, whose company’s ChatGPT models are among the most popular in the United States, told Axios in an interview earlier this year. In September, Anthropic CEO Dario Amodei argued that AI companies should “slow the pace at which we improve the capabilities of AI models” due to possible threats that emerge during rapid development.
“Left unchecked, it could outrun our ability to understand and control these systems, and so must be pursued very carefully, if at all,” Amodei wrote in an essay at the time.
Steven Richards
Source: https://justthenews.com/politics-policy/cybersecurity/warnings-ai-powered-cyber-attacks-and-autonomous-hacking-reach-fever
No comments:
Post a Comment