Cybersecurity firm CrowdStrike has issued a warning regarding a rise in cyberattacks originating from China, aimed at stealing artificial intelligence technology to bridge the technical divide with the United States.
According to a recent report, Chinese-linked entities were responsible for over 58% of state-sponsored targeted attacks against technology firms, with a primary focus on acquiring AI assets.
“China-nexus adversaries are escalating espionage against technology organizations to steal the AI capabilities and intellectual property they cannot build fast enough on their own,” CrowdStrike stated.
The analysis, which covers the period ending March 31, suggests that U.S. restrictions on high-end AI training chips have hindered Beijing’s domestic development, prompting the development of homegrown models designed to lower operating costs while attempting to maintain competitive performance.
CrowdStrike further noted that Chinese-affiliated actors targeted government communications in Southeast Asia and leveraged vulnerabilities to maintain persistent access to North American technology organizations.
The Cyberspace Administration of China did not immediately respond to requests for comment.
This trend follows complaints from AI leaders Anthropic and OpenAI earlier this year, who claimed that Chinese firms were extracting competitive intelligence. At the time, some analysts noted that the line between competitive research and illicit behavior can often be ambiguous.
In response to these threats, Anthropic has recently highlighted the robust cyber capabilities of its newest Mythos model, which has been deployed to CrowdStrike and other partners. Anthropic also released a public version of the model, Claude Fable 5, which the firm Artificial Analysis ranks significantly ahead of other leading models.
Additionally, CrowdStrike identified efforts by North Korean-affiliated entities to infiltrate IT workforces across Asia, Europe, and North America, primarily as a means of generating revenue for the regime.