10
Loading video...
By Chris Matthews
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
Hackers are using artificial intelligence to find weaknesses in computer systems and launch attacks with little or no human intervention, Google has warned.
The technology giant said Monday that criminals are deploying AI systems capable of identifying security flaws, developing malicious software and attempting to exploit vulnerabilities automatically.
In its latest AI Threat Tracker report, Google's Threat Intelligence Group revealed that researchers had uncovered a system designed to exploit security weaknesses before companies had time to install updates.
Read More on News
In a statement accompanying the report, Google said: "The era of autonomous attacks has arrived."

...
How Are Hackers Using AI?
Google said researchers discovered an exposed server hosting an AI-powered hacking system known as exploit_agent.
The software was designed to examine updates released by developers, identify which security weaknesses had been fixed and create attacks targeting computers that had not yet been updated.
It could then attempt to exploit those weaknesses without a hacker directing each step.
Researchers also found data stolen from victims on the server, alongside tools for searching compromised networks, breaking into accounts and installing malicious software.
Evidence suggested the operator was particularly interested in French websites associated with the media, entertainment and education sectors.
Google did not disclose how many organizations had been successfully attacked using the system.
Another AI-powered framework, known as CAIRN, was observed in intrusion activity targeting organizations including those in the U.S. aerospace sector.
A separate tool, based on a system called PentAGI, was used to scan organizations across Asia for weaknesses in their computer systems.
Rather than spending hours manually searching for vulnerabilities, criminals can use these systems to automate parts of the process and respond to what they find.
Google warned that the technology could make cyberattacks faster and more accessible to criminals with limited technical expertise.
What Does It Mean for Cybersecurity?
Hackers have long exploited outdated software and stolen passwords, but AI is making it possible to carry out some of those attacks much faster.
John Hultquist, chief analyst at Google's Threat Intelligence Group, said: "For the most part, attackers aren’t using new groundbreaking methods to break into systems—they are still going after known, unfixed issues, just with much greater speed and scale."
He warned that security teams would need to make greater use of automated defenses to keep pace.
Google also identified criminals stealing access to paid AI services and selling it to other users.
In some cases, attackers gained access to companies' cloud accounts and used their computing resources to run AI systems, leaving the businesses facing substantial charges.

...
The company said some victims faced more than $100,000 in fraudulent charges in a single day, while cumulative costs reached $1 million in the most serious cases.
The report also detailed how North Korean-linked hackers and other groups were incorporating AI tools into their operations, including efforts to develop malware and compromise computer systems.
Separately, Google identified state-backed groups and other operators using AI to create fake online identities and produce political content intended to influence public opinion.
Google said autonomous hacking tools still had weaknesses. Their activities often left recognizable traces that security teams could identify, while some AI-generated exploits failed to work.
The company expects the technology to improve, making attacks harder to detect and reducing the need for human involvement.
Google said it was developing its own AI-powered security tools to identify and repair vulnerabilities before criminals could exploit them.
Contact Newsweek editors for this story: Robert Greenall and Rebecca Flood
View 0 Comments
Submit Correction
View Editorial & AI Guidelines