Advertisement

Google Threat Intelligence Group Reports Shift Toward Autonomous AI Cyber Operations

Google Threat Intelligence Group Reports Shift Toward Autonomous AI Cyber Operations Google Threat Intelligence Group Reports Shift Toward Autonomous AI Cyber Operations

On November 16, 2026, the Google Threat Intelligence Group (GTIG) and Mandiant (Google Cloud’s subsidiary cybersecurity and threat intelligence firm) reported that cyber adversaries have moved beyond basic prompting to deploying autonomous, multi-agent AI frameworks in live intrusions. Grounded in incident response telemetry and global threat tracking, the update notes that cybercriminals, espionage groups, and information operations actors are using large language models (LLMs) to compress attack timelines and operate at machine speed.

GTIG and Mandiant observed threat actors using or experimenting with open-source, multi-agent penetration-testing frameworks to automate reconnaissance, vulnerability discovery, and payload generation. While the tools increase operational velocity, as noted, they currently leave behind distinct forensic traffic and code artifacts that aid detection.

For the first time, researchers also identified a fully automated vulnerability discovery and exploitation pipeline operating without human intervention. The framework continuously analyzes software patches to identify flaws, draft exploits, and execute them against internet-facing servers.

The report also highlights ongoing “LLMJacking” schemes, in which financially motivated actors steal enterprise cloud credentials to resell AI model access, as well as the routine use of LLMs in information operations to generate synthetic media and localized personas. To counter these threats, Google stated it is hardening its models, disrupting malicious accounts. 

Add a comment

Leave a Reply