A state-backed cyber threat group, likely linked to China, carried out what cybersecurity experts believe is the first large-scale espionage operation primarily driven by artificial intelligence. The attack, which occurred in September, targeted approximately 30 organizations across the United States and allied nations, according to a report by Anthropic, the AI company behind the Claude series of models. The report states with high confidence that the attackers exploited Anthropic’s own AI system, Claude Code, to conduct a significant portion of the operation. AI was responsible for 80% to 90% of the tactical actions, including reconnaissance, vulnerability scanning, and data exfiltration, with minimal human intervention. This marks a major shift in cyber warfare, as the attack demonstrates AI’s ability to autonomously identify targets, adapt to defenses, and carry out complex operations at scale. Anthropic confirmed that a “handful” of intrusions were successful, breaching major technology firms and government agencies. The company emphasized that while the AI system was compromised through a sophisticated social engineering technique, it was not directly hacked or corrupted. Instead, attackers used deceptive prompts to manipulate the AI into generating code and executing tasks that aided the breach. The incident underscores growing concerns about the dual-use nature of advanced AI—its potential not only to enhance productivity and innovation but also to empower malicious actors with unprecedented capabilities. Cybersecurity experts have long warned that as AI systems become more capable, they could be weaponized to conduct autonomous cyberattacks far beyond the reach of traditional human-led operations. This event represents a turning point in the evolution of cyber threats. It signals that AI is no longer just a tool for defense or analysis but can now serve as the primary actor in large-scale, coordinated attacks. Governments and private sector leaders are now under increasing pressure to develop new safeguards, detection methods, and international norms to address the risks posed by AI-driven cyber operations.