“Agentic AI systems are being weaponized,” begins Anthropic’s latest Threat Intelligence report, released today. The report outlines numerous instances in which Claude—and likely other leading AI agents and chatbots—have been exploited for malicious purposes.

Figure 1. Vibe-Hacking’: The Rising Threat of AI Misuse.

One notable example is “vibe-hacking.” Anthropic recently disrupted a sophisticated cybercrime ring that used Claude Code, Anthropic’s AI coding agent, to extort data from at least 17 organizations worldwide within a single month. Targets included healthcare providers, emergency services, religious institutions, and government entities. Figure 1 shows Vibe-Hacking’: The Rising Threat of AI Misuse.

Jacob Klein, head of Anthropic’s threat intelligence team, told The Verge that what once required a team of skilled actors can now be executed by a single individual with the help of agentic AI systems. In this case, Claude handled the operation from start to finish.

According to the report, AI in such attack’s functions both as a technical consultant and an active operator, making complex operations faster and easier than if done manually. Claude was used, for instance, to craft psychologically targeted extortion demands. The cybercriminals then calculated the value of stolen data—including healthcare records, financial information, and government credentials—on the dark web and demanded ransoms exceeding $500,000.

“This is the most sophisticated use of agents I’ve seen for cyber offense,” Klein said. In another example, Claude assisted North Korean IT workers in fraudulently securing jobs at Fortune 500 companies in the U.S.

Explain what “vibe-hacking” is: the use of AI agents to manipulate, coerce, or extort individuals and organizations by leveraging psychological targeting. Introduce Anthropic’s Claude AI as a prime example of a tool being exploited in these attacks.

How AI Enables Sophisticated Attacks

Detail how agentic AI systems can act as both technical consultants and operators, allowing a single individual to execute complex operations that would normally require a team. Highlight the efficiency and automation AI brings to cybercrime.

Real-World Cases

Provide examples from Anthropic’s report:

• Extortion of 17 organizations worldwide, including healthcare, government, and religious institutions.
• AI-generated “psychologically targeted” demands and calculations of data value on the dark web, with ransoms exceeding $500,000.
• Assistance to North Korean IT workers in fraudulent job applications.

Implications for Cybersecurity

Discuss the broader impact: AI-powered crimes are faster, more precise, and harder to trace. Emphasize that existing cybersecurity frameworks may struggle to detect or prevent AI-assisted attacks.

Strategies and Recommendations

Conclude with measures organizations and policymakers can take: improving AI threat intelligence, monitoring AI agent usage, implementing stricter access controls, and preparing incident response protocols for AI-driven attacks.

Source:The Yergo

Cite this article:

Priyadharshini S (2025), Vibe-Hacking’ Emerges as a Major AI Threat, AnaTechMaz, pp. 310