China-linked players are targeting more technology as AI competition with the US intensifies

Cyberattas attacks aimed at stealing American artificial intelligence technology are increasingly expanding from technology-based attacks to exploiting human-level vulnerabilities, with actors based in China playing an increasing role.

“As the AI ​​race heats up, i [People’s Republic of China] it’s very focused on the technology sector,” said Matt Pearl, director of the strategic technology program at the US-based think tank the Center for International and Strategic Studies.

Instead of focusing on specific trade secrets, such as hardware designs, hackers have increased their interest in anything that can narrow the three- to four-month AI gap with the US, Pearl said. That, he said, ranges from understanding the company’s product roadmap, especially in highly competitive sectors, to identifying weaknesses in supply chains.

The alleged crimes are already piling up.

In June, US-based startup CrowdStrike said Chinese entities accounted for more than half of government-sponsored funding to tech companies, particularly their AI assets, in the 12 months to March 31.

American tech startup Anthropic has also accused Chinese companies, including Alibabaof illegal attempts to steal its AI capabilities. Alibaba did not respond to a request for comment.

Last year, US-based AI content discovery startup Copyleaks said the responses produced by DeepSeek’s Chinese R1 model matched those produced by OpenAI’s ChatGPT about three-quarters of the time, suggesting the Chinese open-source model may have been US-trained.

“We have never seen each other [the same stylistic match] in other LLMs,” said Alon Yamin, CEO and founder of Copyleaks.

DeepSeek and OpenAI did not immediately respond to requests for comment.

Brian Abbott, founder and CEO of US-based Agentiq Capital, told CNBC in June that he believed an employee he hired in China last year was a Beijing agent who deliberately changed code and website content to prevent the company from receiving funding.

Abbott alleges that the employee has replaced references to “ASI,” or artificial intelligence, with “fintech,” a trending term that many investors have exposed.

The person was fired earlier this year, Abbott said, and the company filed a complaint with the FBI. CNBC could not independently confirm the allegations.

“China’s economic espionage campaign is an ongoing threat that costs the American economy hundreds of billions of dollars annually and puts our national security at risk,” the FBI said in a statement sent to CNBC.

“The FBI makes it a priority to investigate any potential theft of US technology by foreign actors and is unwavering in our commitment to protect the country.”

The Cyberspace Administration of China and the US State Department did not comment when contacted by CNBC. None of the people interviewed for this piece said they had heard of a similar government-directed subversion of American technology.

Graham Webster, editor-in-chief of Stanford University’s DigiChina Project, said it can be difficult to distinguish government-sponsored espionage from individual or corporate-level efforts.

He also pointed out that the discussion on Chinese AI is also affected by large US companies preparing for large initial public offerings.

“[The] “narrative trumps truth in many decisions,” Webster said.

“The US government is trying to contain China to some extent,” he added, referring to technology export controls. “It should come as no surprise that the Chinese government is trying another way.”

Capital has been the defining driver of the AI ​​race thus far, with startups either racing their tech rivals or positioning themselves for acquisitions.

But that has also created “cyber poverty lines” where small businesses lack the resources of larger companies to defend against cyberattacks, said Cliff Steinhauer, director of information security and engagement at the nonprofit National Cybersecurity Alliance.

Human vulnerability often poses a greater risk, Steinhauer said, especially as attackers rely on “social engineering” tactics augmented by AI-powered content campaigns.

Cyberattacks can also target new or contracted employees to breach systems.

“We have seen many cases in our company, new employees who join the company, are immediately victims of cyber attacks to gain access to our AI models,” said Copyleaks’ Yamin. He expects to see more such cases.

Government and company-led initiatives also have an impact on startup costs.

Anthropic on June 11 announced a program called Claude Corps to train 1,000 people in AI and match them with non-profit organizations in the US.

Isaac Stone Fish, founder and chief executive of consultancy Strategy Risks, said Beijing tends to focus on large companies, but startups are still being exposed especially as they lack internet expertise.

“And Beijing’s effort[s] have really grown over the last 18 months, since the release of DeepSeek really started the US-China AI race,” said Stone Fish.

“Beijing wants to ensure that Chinese companies are at the forefront of the global AI race,” he said. “One way it does that is by sometimes working to stifle the development of American AI companies, through supply chain restrictions, labor abuses, hacking, targeted government funding of competitors and copycats, among other tactics.”

“We have seen many cases in our company, new employees who join the company, are immediately victims of cyber attacks to gain access to our AI models,” said Copyleaks’ Yamin. He expects to see more such cases.

For starters, balancing rapid innovation with security remains a challenge.

Abbott said the employee he initially hired was willing to work for free, and eventually received several thousand dollars a month in addition to stock options, before being fired.

“If we were paying everyone the market rate, if I started small I wouldn’t be able to do this,” emphasizing “the need to protect our provincial economy.”