Anthropic AI Used to Automate Data Extortion Campaign

Anthropic revealed that a cybercriminal group tracked as GTG-2002 abused its Claude Code tool to automate a large-scale data theft and extortion campaign, marking a new evolution in AI-enabled cybercrime. In its August threat intelligence report, Anthropic said GTG-2002 made wide use of Claude Code to conduct automated reconnaissance, intrusions, credential harvesting, data exfiltration, and even the creation of malware and anti-detection tools.

The operation reportedly targeted at least 17 organisations worldwide and involved real-time assistance with network penetrations, privilege escalation, and lateral movement, with ransom demands sometimes exceeding $500,000. Anthropic noted that Claude Code helped analyse exfiltrated financial data to set ransom amounts and even generated HTML ransom notes displayed on victims’ machines.

The company said GTG-2002 did not encrypt data; instead, the activity centred on data theft and extortion, and Anthropic described the case as indicating a shift to “vibe hacking,” where agentic AI performs attacks as well as advises.