Last week Fabian J.G. Westerheide, CEO of AI for Humans and noted expert on artificial intelligence (AI) strategy, in writing for Forbes, called China “the first AI superpower.” Westerheide noted that “the People’s Republic of China has the most ambitious AI strategy of all nations and provides the most resources worldwide for its implementation.”
He suggested that China’s strategy on AI is even bigger than its “Made in China 2025” plan and could be linked to a new (digital) Silk Road, helping China become the world’s largest economic power.
Westerheide is not alone in this assessment of Chinese AI capabilities.
According to a study from research firm Global Data’s defense technology writer Harry Lye, “whoever achieves AI proliferation first will be leagues ahead of the competition, adversary or ally. AI development has emerged as the new arms race, but this time with a much more advanced toolkit. The stakes in this race are higher than ever, but also misunderstood.”
Lye added that world powers are still judged by their respective fleets of tanks and aircraft carriers – also a focus for China – rather than unmanned systems.
For a nation that hasn’t always been forthcoming, China has actually been quite open about its AI intentions. In a 2018 session of the Politburo Chinese Premier Xi Jingping said the nation must “ensure that our country marches in the front ranks where it comes to theoretical research in this important area of AI, and occupies the high ground in critical and core technologies.”
The United States military has taken notice, and U.S. Secretary of Defense Mark Esper spoke on the dangers of allowing China to become the dominant power in AI. Esper, speaking at the National Security Commission on Artificial Intelligence Public Conference last November, said, “The Chinese People’s Liberation Army is moving aggressively to deploy (AI) across many warfighting domains. While the U.S. faces a mighty task in transitioning the world’s most advanced military to new AI-enabled systems, China believes it can leapfrog our current technology and go straight to the next generation. We have to get there first.”
The AI Threat Considered
To understand whether the threat of an AI arms race is likely, it is first necessary to understand the state of AI.
“AI is an unfortunate term, one that was coined in 1956 and has taken on a lot of different meanings,” explained Tom Stefanick, visiting fellow for foreign policy at the Brookings Institute. “With AI there is a lot of different things and it can get confusing – but we need to understand it isn’t the great threat that some visionaries and futurists suggest it will be.”
Simply put AI is about autonomy – software algorithms that can solve problems as only software can do.
“We also have to think about what we mean by an arms race, and with nuclear arms there was the very real threat of mutually assured destruction,” Stefanick told ClearanceJobs.
AI is just one part of a much greater issue of computer development.
“AI goes hand and hand with the development of quantum computing,” said Brad Curran, industry principal for aerospace, defense and security at Frost & Sullivan.
“Whoever gets there first will have a huge advantage, and the same goes with the development of laser weapons,” Curran told ClearanceJobs. “We want that technical dominance, but right now it is hard to see inside the Chinese defense establishment to really know how far along they are.”
Curran added that the threat of Chinese dominance in AI, and even its ability to “leapfrog” the United States in AI development could be overstated. The People’s Republic of China has played catch up with the West since its earliest days in the late 1940s, and for much of its early existence it relied on Soviet military hardware, either directly or by copying the Soviets.
China is only now making strides in conventional weapons, and where it has, it still trails the United States military.
“Building an aircraft carrier is one thing, but having the capability like ours to use it on a global stage is another step that China isn’t close to taking,” added Curran. “They still copy Russian designs, and they also hacked our defense networks so they copied our gear and technology. We need to take them seriously, but with conventional weapons as well as in AI they are a long ways from being a peer.
Alphago Zero-Sum Game
The other part of understanding the AI threat may come down to how it is really simply about problem solving. This was best highlighted in a computerized version of the classic Chinese board game Go, where a computer program – Alphago Zero – learned to master it. It was not taught the moves, but after millions of moves was proven to be nearly unbeatable.
This is the advantage AI could offer.
“This is really a game, and there is both offensive and defensive in play and it is mixed where you are always switching between the offensive and defensive, almost like you are making two moves at a time,” said Stefanick.
Where AI fits in is that it can make all of this a little easier to conduct. Even if China were to make massive leaps in the development of AI or machine learning, that may not immediately give it a major advantage.
“There is the ability to break one’s encryption, which is a real consequential thing, but on the other side are the efforts to further secure that encryption,” Stefanick noted. “These are very symmetric efforts.”
The other consideration is that while AI could give a nation a leg up, major powers no more want to instigate a cyber war than a nuclear war. China is as vulnerable as any other nation, possibly more so due to how its infrastructure is spread out.
“Right now we see AI framed as part of an arms race, as who has more research level papers and who has more studies,” Stefanick told ClearanceJobs. “But you have to see that this is just the top of the pyramid. It is built upon the data, the people labeling and analyzing the data, the technicians and the software people. Even the best AI can’t move forward without what is building it up.”
In other words, the layers of important skill sets is the real issue – not who may be winning the development of the most powerful AI.
“We have a shortage of skill sets that need to be addressed in the U.S.,” warned Stefanick. “It isn’t just IT and cybersecurity, but those of technical training as well. A lot of engineers come out of school with the knowledge but they lack the skills to build something. And AI isn’t going to solve that problem.”
