China is militarizing AI, and that’s the AI threat we should worry about

Like Schrödinger’s fabled cat, AI appears to live in a state of superposition in our zeitgeist — it can be thought of as both humanity’s salvation and damnation. This paradox has states, Congress, and the Biden administration bending over backwards to pass laws and promulgate rules to quell, for the most part, hypothetical harms.

As of February 2024, there were 407 “AI-related” bills in state legislatures alone. Worse, these bills’ legislative focuses are all over the place, ranging from privacy to digital equity. What’s more, President Biden issued an executive order that requires dozens of agencies to develop myriad AI standards to do everything from advancing “equity and civil rights” to promoting public safety.

The irony of this shotgun approach to AI is that we run the risk of missing the imminent threat of AI altogether — China’s quest to militarize the technology. By dividing our focus on less immediate harms, we are giving up the AI ghost to China, for as we doddle over its theoretical applications, China is building military-grade AI machines.

The Chinese Communist Party isn’t shy about its ambitions. As the Defense Department has explained, “China’s efforts include developing and incorporating military-AI…to build an ‘intelligentized’ force…to wage and win dynamic wars.”

What do these dynamic wars look like? According to scholars at the National Defense University in Washington, D.C., they include “intrusions by hackers for the purpose of stealing information, attempts by a country to paralyze another country’s infrastructure through software bugs, and misinformation spread via the Internet.”

This is why China has used its own companies’ technologies to flood global markets to obtain access points into those political systems. 

Its invasion has already started in the U.S. by strategically implementing its tech up and down our Internet stack. Since at least 2010, China has attempted to deploy its telecom networks, courtesy of Huawei and ZTE, all over the U.S. to extract data. To conduct citizen surveillance and espionage, the CCP has solicited the help of DJI Drones’ mobile app and TikTok to be placed on millions of mobile devices. It has also deployed millions of cheap routers via TP-Link and Jetstream, which have noted vulnerabilities that China-based hackers can use to hack into American home networks.

So how do we reorient our focus to combat that threat? We cut off the dataflows to China from the U.S.

Thankfully, we have enacted bipartisan measures to do just that. The Federal Communications Commission’s rip-and-replace program has been removing Huawei and ZTE equipment from our networks. In a similar vein, FCC Chairwoman Jessica Rosenworcel and Commissioner Brendan Carr are closing a loophole that bad-faith actors, like Huawei, could exploit to sneak their banned equipment into our networks. 

On the application side, in 2022, the U.S. banned the use of TikTok on government devices. Even better, the bipartisan “Protecting Americans from Foreign Adversary Controlled Applications Act” is now law of the land. This law creates a framework for the U.S. to protect Americans from commercial apps that are owned and controlled by governments of foreign adversaries — specifically Iran, China, North Korea and Russia.

Better yet, we are not limiting our scope to China. The Commerce Department is using its Information and Communications Technology and Services regulations to ban sales by Kaspersky Lab — a Russian antivirus vendor — for the national security threat it poses.

All of these steps, although incremental, are vital to cutting off dataflows to China and other foreign adversaries.

But more must be done. For instance, the government has only provided the FCC with enough resources to remove 40 percent of the Chinese equipment in our networks. Even with the rip-and-replace program in place, there are at least 24,000 pieces of equipment from Huawei, ZTE and other Chinese companies spread over 8,000 locations in our wireless telecom networks. Congress needs to provide more funding to finish the job we started in 2019.

Worse, in 2021 and 2022, the U.S. Secret Service purchased Chinese drones that were known to have spyware. U.S. military bases purchased smart TVs from Chinese manufacturers TCL and Hisense that an investigation revealed were transmitting information back to China. We need to remove these devices immediately.

Unbelievably, there are at least 150 million Chinese routers in our global networks; some may even be used by our own government. Think about the implications of that. We now know that routers allow hackers to peer through walls by using Wi-Fi signals to create de facto heat maps and show where people are throughout the buildings, or even map out classified areas. Imagine these routers are placed in the Pentagon or military bases, which some are. It’s why Congress needs to pass the ROUTERS Act, so that we can effectively target investigations to triage that threat.

If data is what fuels AI systems, then we need to close every pump the Chinese government can access. That is why Congress should put its AI focus here where it’s needed instead of trying to herd AI cats.

Joel Thayer is president and board member of the Digital Progress Institute.