Google warns that fully-automated AI attacks are coming soon
by Bill Toulas · BleepingComputerThreat actors are now relying on agentic workflows powered by AI-based offensive frameworks to automate vulnerability discovery and exploitation, accelerating attacks to machine speed.
Although AI was initially used experimentally in the attack flow, Google Threat Intelligence Group (GTIG) has observed its large-scale adoption by both financially motivated and state-sponsored threat groups.
A notable example is the use of CAIRN, an open-source multi-agent penetration-testing framework, in an attack against the U.S. aerospace sector, and suspected North Korean hackers using Claude Code to deliver malware.
The researchers estimate that in the next three months threat actors will be able to launch unsophisticated attacks that require "little to no human interaction" and rely mostly on multiple AI agents.
Their projection is that "sophisticated, fully autonomous" agentic attacks are likely to emerge in the next six months.
Source: Google
In an advisory last week, the FBI warned that multiple threat actors started weaponizing the ARTEX AI pentesting framework since August and launched attacks against organizations in the financial, government, technology, defense, healthcare, and other sectors.
Using AI to find and exploit flaws
Google says attackers now systematically abuse open-source pentesting suites to run fully automated attacks, including ARTEX, CAIRN, CyberStrikeAI, and HexStrike AI.
Attackers used these frameworks to carry out adaptive attacks against internet-exposed endpoints vulnerable to low-complexity flaws for which large language models (LLMs) could easily build exploits.
In one example GTIG highlights in its AI Threat Tracker report, hackers used a multi-agent system called ‘exploit_agent’ to fully automate vulnerability finding and exploitation in attacks targeting French media, entertainment, and education websites.
“In the past quarter, GTIG identified threat actors using LLM frameworks to fully automate vulnerability discovery and exploitation without a human in the loop for the first time,” Google says.
“More specifically, GTIG identified an exposed staging server hosting a multi-agent vulnerability research and weaponization framework, dubbed exploit_agent, that integrates with multiple frontier and local LLMs.”
Google discovered this framework from the attacker’s exposed staging server, which also included stolen victim data, reconnaissance information, webshells, and brute-forcing scripts.
The system analyzes software updates on popular web apps to find signs of security flaws and automatically attempts to exploit them.
Google warns that the deployment of such frameworks in attacks has dramatically decreased the time defenders have to patch their systems, while it also produces false leads.
The researchers noted the recent example of PHP2Shell, where hundreds of purported proof-of-concept (PoC) exploits, mostly AI-generated, appeared shortly after PHP announced the two flaws. This added a significant burden on defenders, as they had to verify the code and develop protections.
Source: Google
Other AI usage trends
Beyond automated attacks, Google spotted multiple attacks from UNC6836, an activity cluster with suspected links to “Midnight Neptune,” targeting cryptocurrency firms in North America and Europe, using modified Claude Code.
After the initial compromise, the attackers modified the Claude Code app's settings so that at launch it triggered the download of the SOMBERMEME backdoor.
Google reports AI is now routinely used in influence operations by state-backed groups, commercial vendors, and political networks.
According to the researchers, LLMjacking has become a widespread problem, with attackers increasingly exploiting this new revenue stream at significant cost to affected organizations.
Uncover email threats your existing defenses may miss.
Adaptive Security's AI agents analyze email intent, context, and behavior to uncover phishing, business email compromise, and invoice fraud.
See what's getting through with a 14-day read-only scan of Microsoft 365 or Google Workspace, with no MX changes.