GTIG AI Threat Tracking Report: Agentic threat activity crosses a critical threshold
GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold | Google Cloud Blog
Google Threat Intelligence Group (GTIG) and Mandiant released a quarterly AI threat tracking report, stating that agentic threat activity has crossed a critical threshold: multiple threat actors used open-source multi-agent attack frameworks such as ARTEX, CAIRN, and PentAGI in real intrusions, and GTIG also said it identified for the first time the exploit_agent framework, which fully autonomously completes vulnerability discovery and weaponization without human intervention, with its targets concentrated on French media, entertainment, and education industry websites. The report counts that from January to August 2026, the overall number of vulnerability disclosures and exploitations nearly doubled, with zero-day exploitation rising from an average of 8 per month in 2025 to an average of 11 per month in 2026, surging to 22 in August. In LLMJacking incidents, fraudulent charges exceeded $100,000 in a single day, with the most severe case accumulating abuse of up to $1 million; stolen resources were resold through underground LLM Mixer storefronts at $0.10 to $25 per million tokens.