Anthropic disrupts malicious AI campaigns from Russia, China

https://live-hls-7agy.livepush.io/live_abr_cdn/nsitWAl1pTZtwca/emnmZbKZcO8RfwJR/index.m3u8

Anthropic has reported significant disruptions to various malicious uses of its Claude models, with a particular focus on cyber threats linked to Russia and China.Over the past eight months, Anthropic identified and neutralised attempts to exploit its AI technology, including a suspected cyber espionage campaign connected to Russia and efforts by Chinese companies to replicate Claude’s capabilities.

AI threats addressed

Anthropic’s Threat Intelligence report indicates that both cybercriminals and state-sponsored hackers are increasingly leveraging AI to facilitate and execute cyberattacks.

The company noted that human operators often play a supervisory role rather than being directly involved in these operations.

Anthropic stated it had thwarted activities from seven labs located in China, including Alibaba, Moonshot, DeepSeek, and Xiaomi.

Reports suggest that Alibaba orchestrated the largest illicit attack aiming to extract Claude’s capabilities to enhance its Qwen models.

During this period, Anthropic detected over 151 million exchanges involving Alibaba, with activity peaking at nearly 3 million daily from over 3,500 fraudulent accounts.

The process of distillation allows smaller AI models to be developed using outputs from larger models to reduce training costs.

Moonshot and DeepSeek are accused of sending live customer interactions, which may have contained sensitive data, through Claude for training purposes.

A group linked to Russia, known as Midnight Blizzard, is alleged to have conducted phishing operations targeting Ukrainian officials using AI throughout its activities.

The group reportedly developed a system that identifies and evades security measures by rewriting its malicious code to avoid detection.

Ongoing security concerns

Anthropic continues to monitor AI-related threats as cyberattacks evolve.

The focus remains on understanding and mitigating the risks associated with malicious use of advanced AI technologies.

As developments unfold, the company emphasizes the importance of vigilance against emerging threats.