Anthropic Unveils AI-Driven Online Espionage Plot, But Meta's AI Expert Disagrees

The American artificial intelligence (AI) developer Anthropic dropped a bomb about a cyber-espionage drive organised by a Chinese state-sponsored group while using its AI model, Claude, to trigger massive portions of a large-scale attack.

In a press release posted on its website, Anthropic explained how it found out about the cyber-espionage plan.

'In mid-September 2025, we detected suspicious activity that later investigation determined to be a highly sophisticated espionage campaign. The attackers used AI's "agentic" capabilities to an unprecedented degree—using AI not just as an advisor, but to execute the cyberattacks themselves,' the AI developer wrote.

It also mentioned that the Chinese state-sponsored group called GTG-1002 exploited their Claude Code tool to try to infiltrate approximately 30 global targets, and succeeded in a few cases.

'The operation targeted large tech companies, financial institutions, chemical manufacturing companies, and government agencies. We believe this is the first documented case of a large-scale cyberattack executed without substantial human intervention,' the company added.

Uncontrolled Hacking with Minimal Supervision

Anthropic's discovery claimed that the AI hacking is more than using codes or tools. This time, AI is used as the key agent of the operation.

In a post shared on X (formerly called Twitter), AI analyst Rohan Paul commended the cyber espionage campaign reveal of Anthropic.

'The AI did 80-90% of the hacking work. Humans only had to intervene 4-6 times per campaign,' he said. Paul also mentioned that using that setup, Claude almost did most of the work, such as mapping networks, locating flaws, writing exploit codes, moving around the systems, unauthorised copying of credentials and staging data.

Paul added: 'Targets included big tech firms, banks, chemical manufacturers, and government agencies, and a handful of intrusions did reach sensitive internal systems before being contained.'

Not Everyone Agrees

Meta's Chief Artificial Intelligence Scientist Yann LeCun turned to X to dispute Anthropic's claim that a Chinese state-sponsored organisation is carrying out the cyber espionage attacks.

According to the French-American AI expert, 'You're being played by people who want regulatory capture. They are scaring everyone with dubious studies so that open-source models are regulated out of existence.'

He also reposted an analysis on X that used Claude's response to check out Anthropic's report.

Based on Claude's reply to the prompt, 'Read this attached paper from Anthropic on a "AI-orchestrated cyber espionage campaign" they claimed was "conducted by a Chinese state-sponsored group." Is there any evidence or proof whatsoever in the paper that it was indeed conducted by a Chinese state-sponsored group? Answer by yes or no and then elaborate.'

The AI said that there is 'no evidence whatsoever to support the attribution to a "Chinese state-sponsored group.'

The response added, 'Anthropic simply asserts that the operation was "conducted by a Chinese state-sponsored group we've designated GTG-1002" without providing any technical indicators, infrastructure analysis, or methodology that led to this attribution.'

Anthropic has yet to reply to what LeCun and all those who disagree with their report. But they continue to stand by what they released and claimed that some of their evidence have yet to be published.