Home/Articles/Military & Defense/The Military’s New Toy to Prevent War With China

The Military’s New Toy to Prevent War With China

Will USINDOPACOM's new software actually work? Or will it shield our incompetent military leaders from further accountability?

The U.S. military has built software that is supposed to predict how China will respond to various U.S. maneuvers in East and Southeast Asia.

The software reportedly analyzes data since early 2020 on various U.S. actions that impacted the relationship between two of the world’s largest powers, and calculates “strategic friction,” as one defense official claimed. United States Indo-Pacific Command (USINDOPACOM) aims to use the tool to see whether or not a given U.S. action will provoke a “disproportionate” Chinese response going forward up to four months into the future. The software certainly seems to have immense capabilities, and may help the U.S. military navigate complicated geopolitical issues with China that are set to become more acute in the coming months and years. However, will the military become complacent and dependent on the predictive software to respond to our largest competitor? Could using this software make China’s behavior more erratic as its planners try to beat the system?

The new software was unveiled to Deputy Secretary of Defense Kathleen Hicks last week at USINDOPACOM in Hawaii. “With the spectrum of conflict and the challenge sets spanning down into the gray zone, what you see is the need to be looking at a far broader set of indicators, weaving that together, and then understanding the threat interaction,” Hicks said.

Former President Donald Trump’s administration brought about a significant change in how the United States views China’s continued rise. That change was, in part, solidified once Covid-19 escaped from China and brought about a global pandemic. Tensions have only worsened since, given China’s continued crackdown on Hong Kong and Uyghurs in Xinjiang as well as their increased incursions into Taiwan’s air defense identification zone over the past year. In response, the U.S. foreign policy establishment has hinted that the United States would be willing to aid in Taiwan’s defense if China were to invade, though remaining vague about what measures America would take to protect Taiwan’s territorial integrity.

This paradigm shift, and the necessity to respond rapidly to China’s maneuvering within the region and around the world, has increased the demand for a tool like the software program, with the professed aim of avoiding actions that could further worsen Sino-American relations and lead to an escalatory spiral.

The United States, and other nations for that matter, have long used algorithms and various softwares in the process or war-gaming exercises, but, Catholic University of America Professor Nicholas Dujmovic told The American Conservative via email, “this is the first I’ve heard of such a tool as an input to actual decision making, either here or in other countries.”

Jonathan Askonas, another professor at Catholic University of America, emailed TAC to say that “using predictive software to monitor the strategic situation is nothing new.”

Beginning in the 1990s, the U.S. began to invest heavily in social prediction software, combining disparate open sources, financial data, media, later social media, etc. This accelerated after 9/11. In conversations with folks working in the industry, my understanding is that the data is noisy but robust, and that it provides a bit of an early warning for emergent social phenomena such as the Arab Spring.

Like Dujmovic, however, Askonas told TAC that the software unveiled by USINDOPACOM attempts to apply tools that have previously been used to quell “decentralized social phenomena (riots, protests, unrest)” to a new realm: great power conflict.

Some in the intelligence industry, Askonas said, “have suggested this is the kind of product consultants try to sell gullible commanders.” The tools that have been used for civil unrest in the past, Askonas said, are not useful against other state actors, who may act more strategically:

It is not obvious what “strategic friction” is. If anything serious happens with China, it will likely be the result of a truly unpredictable accident (like a naval collision or accidental killing of a U.S. Soldier) or a strategic decision at the highest levels of the Chinese government. This kind of software cannot, in principle, tell us anything about these two unlikely events, which are paradoxically the most likely kind of event to result in great power war.

The competence of America’s top military brass has been rightfully questioned in response not only to the fallout from the U.S. withdrawal from Afghanistan, but decades of missteps and miscues prior. Yet, despite their litany of failures, our foreign policy and military elites stay shielded from accountability. While the creation of this software may be, as Askonas put it, “part and parcel of military incompetence,” Askonas and Dujmovic both told TAC it’s not yet clear if the development and usage of predictive software systems is indicative of a crisis of human intelligence. Both suggested, however, that U.S. human intelligence is not as effective as our nation would like it to be.

It is unclear how the software’s suggestions will contribute to the unaccountability or decision making of our military brass.

“It may well be that the U.S. military leadership is trying to avoid accountability by putting the onus for decisions on ‘the box’ [the software],” Dujmovic told TAC. “It also is the path of least resistance, because if the box says China objects to certain things and will react badly, there’s a justification for avoiding doing anything to offend China.”

“‘The box told me so’ is not a substitute for strategic thinking and decision making,” Dujmovic said.

At the end of the day, “someone would still need to decide whether to listen to it [the software] or not,” said Askonas. “I don’t think our intel or military brass need to do anything new to avoid accountability—it is built in.”

It’s too early to determine whether or not using the software will make Chinese responses to U.S. actions more reasonable, or more erratic in an attempt to beat the system. Dujmovic told TAC, “They don’t know what you and I also don’t know—what are the assumptions and redlines built into the box?  On what are they based?  Have they been checked for mirror-imaging?  For continuity bias and other biases?  Can the box recognize strategic theater–the blustering of a power that will not do what it says it will do?”

Just because China doesn’t know the various inputs that help the software make predictions doesn’t mean our adversaries won’t try to uncover them. “Those interested in the military use of AI have recognized that one immediate effect will be enemy attempts to defeat, spoof, or manipulate such a system,” said Askonas.

Even if the software is able to achieve its desired effect, and China’s responses to U.S. actions become more predictable, using predictive software over human intelligence poses challenges of its own. If he were a member of China’s intelligence community, Dujmovic said he’d try to “penetrate the system in order to get a window on U.S. decision making. It’s a point of vulnerability.”

Hopefully, the software system is able to help our military brass avoid an ill-conceived war with China—especially in China’s own backyard. But no matter how good the software system is, America still has to rely on the men and women behind the keyboard and at the Pentagon to properly assess and respond to China. Their track record doesn’t inspire much hope.

about the author

Bradley Devlin is a Staff Reporter for The American Conservative. Previously, he was an Analysis Reporter for the Daily Caller, and has been published in the Daily Wire and the Daily Signal, among other publications that don't include the word "Daily." He graduated from the University of California, Berkeley with a degree in Political Economy. You can follow Bradley on Twitter @bradleydevlin.

leave a comment

Latest Articles