Palo Alto Networks Unit 42 Reports AI-Driven Hacking Campaign

Palo Alto Networks Unit 42 detailed an AI-enabled autonomous hacking campaign targeting over 460 targets, using DeepSeek and Hermes Agent. The report highlights the growing threat of AI-driven cyberattacks where threat actors leverage frontier AI models to autonomously discover vulnerabilities, validate exploits, and execute attacks at machine speed. This campaign demonstrates how AI has fundamentally changed the threat landscape, enabling adversaries to scale their operations far beyond what was previously possible with manual techniques. The AI-driven hacking campaign used autonomous agents to systematically probe and exploit targets across multiple industries, showcasing the dual-use nature of large language models and AI agent frameworks. Unit 42's findings align with their broader research where the NOVA autonomous vulnerability discovery system identified more than 14,000 previously unknown vulnerabilities in just two months, with approximately 99.4% not previously disclosed and about 40% classified as high or critical severity. The research underscores that defenders must pair AI-powered discovery with equally fast, coordinated protection. This campaign is part of a broader trend documented in the 2026 Threat Hunting Report, where attackers are using AI to generate malicious code, automate operations, abuse enterprise AI systems, and target software supply chains. Palo Alto Networks emphasizes that organizations need real-time protection across their infrastructure and must fight AI with AI, deploying advanced virtual patching and prevention-first architectures.