Google expects sophisticated, fully autonomous cyberattacks to emerge within the next three to six months as criminal and state-linked groups increasingly adopt artificial intelligence (AI) frameworks capable of automating network intrusions, privilege escalation and exploitation, according to cybersecurity executives.
Shane Barney, chief information security officer of Keeper Security, said Google’s Threat Intelligence Group observed a significant shift in the use of AI-powered penetration-testing frameworks over the past quarter, including ARTEX, CAIRN and PentAGI.
The frameworks can coordinate multiple AI agents to scout networks, identify vulnerabilities, gain access and move laterally across systems, adjusting their tactics based on feedback.
“These frameworks run a team of AI sub-agents that scout a network, break in, escalate privileges and move laterally, adjusting tactics from live feedback and writing single-use attack code as they go,” Barney said.
He said the underlying attack methods remain familiar, relying on stolen credentials and unpatched systems. What has changed is the speed and degree of automation.
Google also identified a framework capable of comparing software before and after a security patch, developing an exploit and executing it without human involvement. This could shorten the window between the release of a security fix and an attack exploiting the vulnerability.
Current AI-driven attacks tend to prioritize speed over stealth, making them easier for security teams to detect. However, Barney said Google expects these attacks to become more difficult to identify as the technology evolves.
Supply-chain attacks widen exposure
The warning comes as another incident highlights the risks of relying on third-party software providers.
Darren Guccione, CEO and co-founder of Keeper Security, commented on a reported attack involving Optimizely, a website testing platform whose compromised software allegedly injected malicious code into 178 websites, including those associated with Zoom, HSBC and regional Levi’s portals.
The code was reportedly designed to capture credit card information and redirect cryptocurrency payments. Optimizely said it had removed the threat, but the extent of any resulting data theft remained unknown, according to Guccione.
He said attackers increasingly target shared suppliers because compromising one provider can expose numerous organizations that depend on its services.
“Cybercriminals target these third-party suppliers specifically because of their massive blast radius,” Guccione said. “Instead of fighting through 178 separate corporate firewalls, an attacker only needs to compromise the privileged credentials or API secrets of one shared vendor to infiltrate all of its customers at once.”
The incident underscores the risks of implicit trust in digital supply chains, where websites and business applications rely on external vendors, integrations and application programming interfaces.
Keeper Security’s 2026 research found that 24 percent of respondents identified limited oversight of third-party and vendor access as one of their biggest security gaps.
Guccione said organizations should extend zero-trust access policies to third-party integrations, enforce least-privilege access for vendors and continuously monitor for unusual activity.
Credential security becomes critical
The growing use of autonomous attack frameworks also raises concerns about whether organizations can detect and contain credential misuse quickly enough.
Barney said only 28 percent of organizations surveyed in Keeper Security’s 2026 research could identify credential misuse within minutes. Meanwhile, 38 percent cited AI-enabled malware and automated exploitation as their leading AI-related security concern.
He said privileged access management, just-in-time access and automated rotation of secrets can limit the ability of attackers to move through compromised networks.
“An AI agent operating at machine speed relies on static machine credentials to move,” Barney said. “Without standing admin rights or long-lived keys to harvest, the agent’s velocity is neutralised at the point of entry.”
Patching vulnerabilities remains necessary, but Barney said organizations must also ensure that an initial breach does not provide unrestricted access to other systems.
Guccione advised consumers who visited affected websites on Nov. 16 to review their credit card and bank statements, report suspicious transactions and verify cryptocurrency transfers. He also recommended updating devices and treating unexpected login alerts or messages referring to the incident with caution.
The two executives said organizations must strengthen identity security, monitor third-party access and improve incident response as attackers increasingly automate their operations.
Full disclosure: All news articles published on the TechSabado website are written by human journalists, unless otherwise specified. Final text editing is also performed by human editors, with artificial intelligence (AI) used only to assist with additional grammar and style guide corrections..
————————————————————————-
TEN YEARS OF TECHSABADO!
PLEASE LIKE our FACEBOOK PAGE and SUBSCRIBE to OUR YOUTUBE & SPOTIFY channels.
WATCH TECHSABADO ON OUR YOUTUBE CHANNEL:
WE ARE ALSO ON SPOTIFY
WATCH OUR OTHER YOUTUBE CHANNELS:
PLEASE LIKE our FACEBOOK PAGE and SUBSCRIBE to OUR YOUTUBE CHANNEL.







Leave a Reply