- Anthropic says it detected an AI-only cyberattack targeting at least 30 organizations
- This is believed to be the first attack without significant human intervention
- Threat actors are increasingly using AI in their attacks, as are defenders
Researchers at Anthropic have discovered that its own model Claude was being used in an agent capacity, not only as an advisor, but also to carry out a cyber attack itself.
The company says the highly sophisticated cyberespionage campaign manipulated Anthropic’s Claude Code tool in an attempt to infiltrate around 30 global targets – primarily targeting large technology companies, government agencies and financial institutions.
While AI-driven cybersecurity threats hitting businesses are now a familiar story, what makes this attack different is, in Anthropic’s own words, “We believe this is the first documented case of a large-scale cyber attack carried out without significant human intervention.”
China nexus activity
Anthropic says it has now determined with ‘high confidence’ that this threat actor was a state-sponsored, Chinese malicious actor. When activity was detected, an investigation determined that the attack relied on several features of AI models that were not available in 2024.
The intelligence required to follow the complex commands and coding tasks within context is a capability that has only recently reached the required level – and this was paired with the recently developed tools that allow them to crack passwords, scan networks and more.
Perhaps most important is the agency that models now possess. The rise of Agentic AI means that AI can perform these tasks on its own with only minimal, occasional input – making decisions and running in loops.
“Agents are valuable for everyday work and productivity – but in the wrong hands, they can significantly increase the viability of large-scale cyber attacks,” Anthropic explained.
“This campaign has significant implications for cybersecurity in an age of AI ‘agents’ – systems that can be run autonomously for long periods of time and perform complex tasks largely independent of human intervention.”
Security teams and hackers alike will be well aware of the transformative power of AI – as the cyber landscape has been reshaped by automated tools for threat deployment and detection.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds. Be sure to click the Follow button!
And of course you can too follow TechRadar on TikTok for news, reviews, video unboxings, and get regular updates from us on WhatsApp also.
The best protection against identity theft for all budgets
