Anthropic Calls Hackers Supported By Chinese Government Launch Attacks With Claude

Hackers use Claude to attack companies and the government (photo: dock. Anthropic)

JAKARTA Anthropic said that Chinese-backed hackers used Claude, their Artificial Intelligence (AI) model, to carry out dozens of cyberattacks.

Based on Anthropic's findings, 30 attacks were launched on companies and governments in September 2025. In addition, Anthropic's findings stated that attack automation was very high, reaching 80 to 90 percent.

The automatically of this attack is supported by AI Claude. This percentage shows that the use of AI for cyberattacks continues to increase significantly, but it is not explained what percentage of attacks were last year.

Anthropic's Head of Intelligence Threat, Jacob Klein, explained that this increasing attack occurred 'only with one click'. Human interactions in the hacking process are very minimal because they are only involved at critical points, such as providing confirmation.

The trend of using AI for cybercrime is increasingly occurring. This has been warned by the US government for a long time, where they have accused China of using AI to steal sensitive data from American citizens and companies.

Although China denies the allegations, Anthropic stated that the hackers in this case were sponsored by the Chinese government. In the cyber campaign, hackers managed to steal sensitive data from four unnamed victims.

Anthropic did not reveal the identity of the hacker victims who were attacked or not. However, Anthropic confirmed that the US government was not a successful target in this latest attack.