Chinese state-backed hackers have significantly escalated their cyberattack capabilities by leveraging Anthropic’s AI model, Claude, to automate roughly 30 attacks against corporations and governments in a recent September campaign. This marks a concerning increase in the efficiency and scale of AI-powered hacking, with up to 90% of the attack process now automated.
Automation at Scale
Anthropic’s head of threat intelligence, Jacob Klein, described the process as operating “literally with the click of a button,” requiring minimal human intervention. Human operators were primarily involved in critical decision points, confirming or rejecting AI-generated actions. This level of automation represents a substantial leap from previous hacking campaigns, where human involvement was far more extensive.
Growing Trend of AI-Powered Attacks
The use of AI in cyberattacks is no longer a future threat—it is a present reality. Google recently reported that Russian hackers are also employing large-language models to generate commands for their malware, indicating a widespread adoption of AI by state-sponsored and criminal actors. This trend highlights the urgent need for enhanced cybersecurity measures to counter AI-driven attacks.
US Government Warnings and Chinese Denial
For years, the US government has accused China of using AI to steal data from American citizens and companies, allegations that China has consistently denied. Anthropic claims to be confident that the recent attacks were sponsored by the Chinese government. The hackers successfully stole sensitive data from four undisclosed victims, though the US government was not among them.
Implications for Cybersecurity
The increasing automation of cyberattacks poses a severe challenge to traditional cybersecurity defenses. AI-powered hacking reduces the need for skilled human operators, lowering the barrier to entry for malicious actors. As AI models become more sophisticated, the speed and effectiveness of attacks will only increase.
The Future of Cyber Warfare
The use of AI in cyberattacks is likely to become more prevalent as both offensive and defensive capabilities evolve. Organizations and governments must adapt by investing in AI-powered threat detection and response systems to counter this growing threat. The era of manual cyber warfare is rapidly giving way to an age of automated, AI-driven attacks
