Beijing escalating AI espionage to catch up with the U.S. on tech, cybersecurity firm says

A prominent U.S. cybersecurity firm, CrowdStrike, has raised alarms about a surge in cyberattacks orchestrated by Chinese entities, with a primary focus on pilfering artificial intelligence technologies. According to their latest report, these Chinese actors were responsible for over 58% of state-sponsored cyber assaults targeting technology firms, particularly those involved in AI development. CrowdStrike emphasized that adversaries linked to China are ramping up their espionage activities against tech organizations in an effort to acquire AI capabilities and intellectual property that they are unable to develop swiftly on their own. This analysis examined incidents occurring over the past year, concluding on March 31. Amid U.S. restrictions that limit China's access to AI training chips, Beijing's technological advancement has faced significant hurdles. Despite these limitations, domestic AI models are increasingly being developed to reduce operational costs while delivering comparable intelligence. The report also highlighted that Chinese-linked cyberattacks have targeted government communication systems in Southeast Asia, while maintaining an ongoing presence within North American tech organizations by exploiting existing vulnerabilities. Notably, the Cyberspace Administration of China did not respond to inquiries regarding these findings. Earlier this year, major U.S. AI firms, including Anthropic and OpenAI, voiced concerns that Chinese companies were extracting sensitive competitive information from American tech leaders. Analysts pointed out that the fine line between legitimate competitive practices and illicit behavior remains murky. In recent developments, Anthropic has been promoting the advanced capabilities of its new AI model, Mythos, which has been made available to CrowdStrike and others. On Tuesday, the company launched a public version called Claude Fable 5, which is reported to be significantly ahead of competitors according to rankings from Artificial Analysis. Additionally, CrowdStrike noted that entities linked to North Korea have also attempted to breach IT teams across North America, Europe, and Asia, predominantly to generate revenue for the regime.