The escalating race for artificial intelligence dominance is becoming a new frontier for cyber warfare, with U.S. cybersecurity firm CrowdStrike issuing a stark warning: China-based actors are intensifying their efforts to pilfer U.S. AI technology, aiming to bridge the perceived development gap. This isn’t merely about acquiring specific trade secrets anymore; the targets have broadened to encompass anything that could accelerate China’s AI capabilities, ranging from product roadmaps to subtle supply chain vulnerabilities.
This trend is underscored by mounting evidence. In a recent report covering the 12 months leading up to March 31, CrowdStrike identified Chinese state-sponsored entities as responsible for over half of the intrusions targeting technology companies, with a particular focus on their invaluable AI assets. This signifies a strategic pivot, moving beyond conventional cyberattacks to exploit human-level vulnerabilities, a tactic amplified by sophisticated AI-driven campaigns.
The scope of these attacks is becoming more nuanced. Rather than focusing solely on proprietary code or hardware designs, malicious actors are casting a wider net. As Matt Pearl, director of the strategic technologies program at the Center for Strategic and International Studies, observed, “As the AI race has heated up, the [People’s Republic of China] has targeted the tech sector increasingly.” This broadened interest extends to understanding a company’s strategic direction, particularly in fiercely competitive sectors, and identifying chinks in the armor of their supply chains.
This evolving threat landscape has already ensnared prominent players. American AI startup Anthropic has publicly accused Chinese entities, including Alibaba, of illicit attempts to acquire its AI capabilities. While Alibaba did not respond to requests for comment, the allegations point to a sophisticated effort to gain an unfair advantage.
Further illustrating this trend, Copyleaks, a U.S.-based AI content detection startup, revealed that responses generated by Chinese startup DeepSeek’s R1 model bore a striking resemblance to those produced by OpenAI’s ChatGPT. This stylistic overlap suggests the possibility that the open-source Chinese model may have been trained on data derived from its U.S. counterpart, a potential violation of intellectual property. Alon Yamin, CEO and co-founder of Copyleaks, noted the unusual nature of this stylistic match, stating, “We haven’t seen [the same stylistic match] in other LLMs.” DeepSeek and OpenAI did not immediately respond to requests for comment.
The personal element of cyber espionage is also coming to light. Brian Abbott, founder and CEO of Agentiq Capital, shared a personal account of an employee hired from China who he believes was a Beijing agent. Abbott alleges that this individual deliberately altered code and website content to hinder the company’s ability to secure venture capital funding, specifically by replacing references to “artificial superintelligence” with the less fashionable term “fintech.” The employee was subsequently dismissed, and the company filed a complaint with the FBI. While CNBC could not independently verify the allegations, the FBI commented that “China’s economic espionage campaign is a continuing threat that costs the American economy hundreds of billions of dollars per year and puts our national security at risk.” The Bureau reiterated its commitment to investigating any potential theft of U.S. technology by foreign actors.
The Cyberspace Administration of China and the U.S. Department of State did not offer comments when approached by CNBC.
Distinguishing between state-sponsored espionage and individual or corporate-level efforts can be a complex challenge. Graham Webster, editor-in-chief of Stanford University’s DigiChina Project, points out that the discourse surrounding Chinese AI is also influenced by the looming initial public offerings of major U.S. tech companies. “The narrative is overtaking reality in a lot of decisions,” Webster remarked. He added, “The U.S. government is trying to hold China back to some extent,” referring to technology export controls, and therefore, “We should not be surprised that the Chinese government tries otherwise.”
**Start-ups Face Heightened Risk**
The intense capital investment fueling the AI race has inadvertently created a divide, with smaller startups often lacking the robust cybersecurity infrastructure of larger corporations. Cliff Steinhauer, director of information security and engagement at the National Cybersecurity Alliance, describes this as a “cyber poverty line.” This vulnerability is exacerbated by the increasing reliance on “social engineering” tactics, which are amplified by AI-powered content campaigns.
New or recently contracted employees are increasingly becoming targets, as attackers seek to exploit these individuals to gain access to sensitive AI models. “We’ve seen a lot of cases within our company, new employees that are joining the company, immediately they’re a target of cyberattacks to get access to our AI models,” Yamin of Copyleaks stated, predicting an increase in such incidents.
Beyond direct cyber threats, government policies also impact the operational landscape for startups. While U.S. companies like Anthropic are launching initiatives to train AI talent, Chinese policymakers are actively supporting their domestic AI sector with benefits such as subsidized computing power and rent-free office space for startups.
Isaac Stone Fish, founder and chief executive of consultancy Strategy Risks, highlights that while Beijing may focus more on large corporations, startups remain particularly vulnerable due to their limited cyber expertise. He noted that Beijing’s efforts have intensified significantly over the past 18 months, coinciding with the release of DeepSeek and the subsequent escalation of the U.S.-China AI race. “Beijing wants to ensure that Chinese companies are at the vanguard of the global AI race,” Stone Fish explained. “One way that it does that is by sometimes working to suppress the development of American AI companies, through supply chain restrictions, employee harassment, hacking, targeted government subsidies of copycat competitors, among other strategies.”
For burgeoning startups, the critical challenge lies in balancing the imperative for rapid innovation with the necessity of robust security measures. Abbott’s experience underscores the financial pressures, where an employee initially willing to work for a nominal amount, including stock options, before dismissal, highlights the cost of competing for talent while simultaneously defending against potential infiltration. “If we paid everybody market rate, for a scrappy start-up I could never afford to do this,” he emphasized, underscoring the “need to secure our economy of start-ups stateside.”
Original article, Author: Tobias. If you wish to reprint this article, please indicate the source:https://aicnbc.com/23324.html
