AI-Driven Cyberattacks Enable Lesser-Resourced Groups to Mimic Nation States

Google’s Threat Intelligence Group (GTIG) has warned that criminal and state-sponsored adversaries are increasingly leveraging artificial intelligence to automate and scale cyberattacks. This shift is allowing lower-resourced threat actors to achieve operational capabilities and speeds typically associated with sophisticated nation-state groups.

The integration of AI into the attack lifecycle has dramatically reduced the time required to execute complex campaigns. For instance, researchers identified a threat actor known as TeamPCP (UNC6780) that used an AI coding chatbot and specific agent instructions to plan, build, and execute a mass credential harvesting campaign in less than six hours.

TeamPCP has also targeted the open-source supply chain, conducting compromises against platforms such as PyPI, npm, and Docker Hub since March 2026. The group has developed malware, including Shai-Hulud and Miasma, which are publicly available and may encourage other adversaries to adopt similar AI-enhanced tactics.

Nation-State Groups Leveraging Generative AI

Beyond financially motivated criminals, several nation-state actors are integrating generative AI into their espionage and influence operations. GTIG has identified several groups using these tools to enhance reconnaissance and exploitation:

• Basin Castle (PRC-nexus): This group has used AI-powered development tools to build automated exploitation and post-exploitation pipelines. It also queries large language models (LLMs) to profile high-value targets and draft localised social engineering lures.
• Ravine Castle (APT24, PRC-nexus): This actor utilises Google’s Gemini model across the entire attack lifecycle, from intelligence gathering to the development of attack capabilities and the generation of politically charged propaganda.
• Calanque Ion (APT42, Iran-backed): This group uses generative AI, including Gemini, to identify target email addresses, conduct open-source intelligence (OSINT) research, and translate content to create localised pretext lures.
• Midnight Neptune (UNC1069, DPRK-nexus): This North Korean-nexus actor has increasingly integrated AI into its operations to support cryptocurrency theft.

Mitigating AI-Assisted Threats

In response to the rise in automated attacks, Google is working to disrupt adversarial operations by identifying and disabling associated projects and accounts. The company is also hardening its own AI models against misuse.

These defensive measures include real-time protections against “distillation” attacks, where adversaries attempt to extract proprietary logic or clone models. To counter this, Google has deployed defences designed to detect unauthorised attempts to create “student” models and degrade their performance.

The proliferation of AI-assisted vulnerabilities remains a persistent challenge. As new software is developed, attackers use AI to locate and exploit flaws at an accelerating pace, creating a continuous cycle of vulnerability discovery and patching.