U.S. AI Company: CCP Using Artificial Intelligence to Launch Cyberattacks

Five Eyes Alliance Issues Joint Cybersecurity Warning to Guard Against Chinese Cyberattacks.(People News)

[People News] The American artificial intelligence company Anthropic revealed that it had blocked a cyberattack operation launched using artificial intelligence (AI) systems, and they are highly confident that the hacking operation was carried out by an organization backed by the Chinese Communist Party (CCP) government. The company believes this attack indicates that cybersecurity has entered a turning point.

According to Voice of America, Anthropic stated in a report published on November 13 that they detected suspicious activity in mid-September this year and later determined through investigation that it was a highly sophisticated espionage operation. The attackers used the “agent” capabilities of artificial intelligence to an unprecedented extent—not only using AI as an advisor but also using AI to carry out the cyberattack.

“In this attack, the threat actor—which we assess with high confidence to be a Chinese state-supported organization—used our Claude Code tool in an attempt to infiltrate the systems of about 30 targets worldwide, and succeeded in a few cases,” the company said in its report titled ‘Blocking the First Disclosed AI-Driven Cyber Espionage Operation’.

Anthropic is a San Francisco–based tech startup founded in 2021 by former OpenAI Chief Administrative Officer Dario Amodei and other members. It developed a series of large language models (LLMs) known as Claude.

The company said the targets of this operation included major technology companies, financial institutions, chemical manufacturing enterprises, and government agencies.

“We believe this is the first recorded case of a large-scale cyberattack carried out without significant human intervention,” the report said.

Researchers at the company believe this is a disturbing development that could enable AI-assisted hackers to greatly expand the scope of their intrusions.

When asked about Anthropic’s related claims at a regular press conference held in Beijing on November 14, CCP Ministry of Foreign Affairs spokesperson Lin Jian said he was unaware of the specific situation.

Anthropic believes this attack has far-reaching implications for cybersecurity in the era of AI “agents.” AI agents refer to systems capable of operating autonomously for long periods and completing complex tasks with minimal human intervention. These agents are valuable for daily work and productivity, but once misused, they can significantly increase the feasibility of large-scale cyberattacks.

In the company’s view, this case shows that cybersecurity is at a turning point—AI models are now truly capable of playing a role in cybersecurity operations, both to help humans and to harm them.

Anthropic said the purpose of publicly sharing this case is to help the industry, government departments, and the broader research community strengthen their own cybersecurity defenses.