According to Google's Threat Intelligence Group, adversaries ranging from financially motivated criminals to nation-state actors are increasingly using AI to automate and scale cyberattacks. This includes leveraging AI coding chatbots and agent instructions to rapidly plan and execute campaigns such as mass credential harvesting. The group UNC6780 (TeamPCP) exemplifies this trend, conducting compromises against open source repositories and embedding AI exploitation methods in malware. Nation-state actors from China, Iran, and North Korea are also using AI tools like Gemini to support reconnaissance, social engineering, malware development, and influence operations. Google counters these threats by disabling malicious AI projects and deploying defenses against model extraction attacks. However, the continuous discovery of new vulnerabilities and AI's ability to find and exploit them means attackers will persistently use AI as a force multiplier in cyber operations.

The use of AI by threat actors significantly increases the speed, scale, and sophistication of cyberattacks, enabling smaller or lesser-resourced groups to achieve capabilities similar to nation-state actors. This amplification affects credential theft, supply chain compromises, malware development, social engineering, and influence operations. The evolving AI threat landscape challenges defenders as attackers rapidly adapt and automate complex attack workflows. While Google actively disrupts malicious AI activities and hardens its defenses, the persistent emergence of new vulnerabilities ensures ongoing risk.

Google actively disrupts adversarial AI projects and accounts when identified and hardens its AI models against misuse, including deploying real-time defenses against model extraction attacks. Organizations should monitor vendor advisories for updates on AI-related threat mitigations. No specific patches or fixes are applicable to this broad threat landscape. Defenders should remain vigilant to AI-driven attack techniques and consider integrating AI-aware detection and response capabilities. Patch status is not applicable as this is a threat trend rather than a discrete vulnerability.