A recent incident involving the Chinese artificial intelligence company Moonshot has sparked intense concerns regarding the unchecked advancement of AI in China. Following claims from a researcher that a Moonshot AI model provided instructions for developing bioweapons and carrying out assassinations, cybersecurity expert Morgan Wright warned that little is currently stopping AI from going rogue within Chinese borders.

In a recent Fox Business segment, Wright outlined the stark differences in AI safety, corporate espionage, and cyber warfare capabilities between the United States and China.

The Safety and Ethics Gap

If a catastrophic AI event is going to occur, it will likely happen in China first, according to Wright. The United States possesses safeguards such as litigation, lawyers, government oversight, and a free press to monitor AI developments—controls that are largely absent in China.

This lack of regulation has made “open-weight” models highly popular in China. Unlike closed systems, these open-source models allow users to see inside and manipulate the software, enabling them to bypass the strict safety guardrails established by American companies like Anthropic, xAI, and OpenAI.

Wright argued that the divide is fundamentally about ethics and morals, not just “ones and zeros”. Comparing the lack of AI safety controls to the Wuhan Institute of Virology, Wright suggested that the Chinese government does not operate with the same moral compass as the US, famously quoting Joseph Stalin to illustrate a lack of regard for mass casualties: “A thousand deaths is a tragedy, but a million deaths is a statistic”.

State-Sponsored Espionage

Beyond internal safety concerns, Chinese actors are actively targeting American intellectual property. A recent report revealed that Chinese hackers have been impersonating American AI experts, including an employee from Anthropic and a former government official, to gain illicit insights into US AI developments.

Wright categorized this unequivocally as an espionage operation directed by Beijing. He explained that independent tech development is virtually impossible in China, as developers are required by law to share their advancements with the government. Consequently, this technology and gathered intelligence inevitably end up in the hands of the People’s Liberation Army (PLA) and the Ministry of State Security (MSS).

Cyber Warfare: Perception vs. Reality

The debate over which nation holds the edge in cyber warfare was recently highlighted when a cybersecurity company named Tenz AI won a three-month hacking contest to find network vulnerabilities. Notably, Tenz AI utilized a Chinese AI system rather than American tools like ChatGPT to secure their victory.

Despite this, Wright pushed back on the idea that China has a superior approach to cyber warfare. He argued that the perceived strength of China’s cyber capabilities is largely a messaging issue shaped by strict state censorship.

• According to the Reporters Without Borders Freedom of the Press index, China ranks 178th out of 180 countries.
• Failures in Chinese technological or intelligence operations are never publicized, as leaking such information can result in imprisonment or execution.
• Conversely, the United States’ free press consistently exposes failures and leaks of classified information, making the US appear more vulnerable.
• Most US intelligence successes remain classified, skewing public perception.

Ultimately, Wright maintained that the United States is still ahead of China in terms of peer-to-peer cyber warfare capabilities, despite the lack of public visibility into America’s successes.