In this summer's Commentary magazine, Arthur Herman writes, "China and Artificial Intelligence: The Cold War We’re Not Fighting."
It describes how much China is investing in artificial intelligence - and for what purposes.
It is the most frightening article I have read in a long time. While it is not specifically about Israel, Israel is affected as well as the entire Western world - and Western civilization.
In short, China is planning to use AI to achieve global hegemony. And the West is not doing nearly enough to stop it.
Even as anti-AI activists in the U.S.—among them some of the technology’s original innovators—were calling for a moratorium on research a year ago, China was paving the way toward an AI-dominated future none of us wants. For the past seven years, China has been moving ahead with its plans to become the world’s AI superpower. This includes building the next high-tech industrial revolution for victory on the battlefield and creating a total surveillance multiverse.China is ramping up AI investment, research, and entrepreneurship on a historic scale. Its generative AI spending is set to reach 33 percent of the world’s AI investment by 2027, up from 4.6 percent in 2022. Those investments will probably reach $13 billion by then, according to a new report from research firm IDC.
What would China do with AI?
Hudson Institute scholar Koichiro Takagi sees the Chinese military’s interest in AI research and applications centering on four main areas (bearing in mind that under Chinese law, anything that private companies develop in AI automatically belongs to the People’s Liberation Army).One area is the autonomy of unmanned weapons, including the development of drone swarms, about which more later.The second is processing large amounts of information through machine learning. For example, the People’s Liberation Army (PLA) is building a network of unmanned weapons and undersea sensors in the waters surrounding China to gather data it can analyze with AI/ML.The third is using AI to speed up military decision-making, including what’s called “strategic reasoning.” AI can sift through multiple options for actions on multiple fronts and domains to arrive at an optimal solution—something that would take a human-led council of war hours, even days, to achieve.Fourth is the military’s interest in cognitive warfare, or actively influencing the brain and neurological systems of their human opponents, to shape the enemy’s will to fight or subdue an opponent without a fight. (The most science-fictional of the four, it is the one about which we have reason to be skeptical, at least for the present.)But probably the most striking and notorious developments within the Chinese AI monolith today are AI’s applications for the total surveillance state.
China is surging ahead on facial recognition, using it to monitor not only Uighurs but every Chinese citizen. Chinese made cameras and communications equipment are used worldwide; they use them to monitor Chinese dissidents but also, potentially, the entire world.
And it gets even worse:
In 2015, then-president of the Academy of Military Medical Sciences He Fuchu insisted that biotechnology will become the new “strategic commanding heights” of national defense. Since then, Fuchu has become vice president of the Academy of Military Sciences, which leads China’s military science enterprise.Zhang Shibo, a retired general and former president of the National Defense University, has named biology as one of seven “new domains of warfare.” In the wake of Covid and the possibility that the virus in the Wuhan lab was being developed for biowarfare purposes, these matters need to be taken very seriously.In all these cases, AI applications can be useful for not only identifying but manipulating and attacking an entire category of persons or groups through targeted viruses and diseases. This is because, at the most basic biological level, DNA itself is nothing more than data—data that can be exploited using AI and machine learning.In that sense, the combination of China’s earlier interest in biotech and its obsession with advancing artificial intelligence may allow China’s military and intelligence services to develop comprehensive digital profiles of specific individuals, nations, and races—a form of high-tech racial profiling that a Himmler or a Mengele might have only dreamed about. By targeting specific weaknesses within a population’s genomic makeup, it might be possible to develop weapons that could harm a specific subpopulation or race.Even more frightening, scientists at the Suzhou Institute of Biomedical Engineering and Technology are using mouse embryos to develop ways to provide key growth information to an AI caretaker, which can then rank the embryos in terms of overall health and genetic potential—enabling researchers to manipulate the growth of embryos to achieve optimal results. In short, the Chinese vision of AI includes a new paradigm for genetic engineering, conducting eugenics on a massive high-tech scale.
This is an unprecedented threat to the West.
What can we do to defend ourselves?
The article falls short on that count, only mentioning generalities, saying "the United States needs to develop an overall AI strategy that aims not just at countering China’s moves in AI but advancing American AI supremacy."
That's good as far as it goes, but it doesn't deal with the problem now.
I am not an AI expert, but it seems to me that there are some tactics that can - and indeed, must - be done today.
AI depends on the quality of the data that it is trained on. I use AI to create most of my cartoons, but the underlying data being out of date means that when I ask for a drawing of an editor speaking to a newsroom, invariably both the editor and the reporters are tie-wearing white males that one would see in 1940s movies about newspapers, and do not resemble newsrooms of today at all.
The AI algorithm works exactly as it should. But its input data is bad.
The old computer adage of "garbage in, garbage out" applies to AI as much as to every other program. To disrupt the Chinese surveillance state, for example, Western state hackers should be breaking into Chinese databases and modifying the oceans of data the AI is using. For example, changing the faces of people the Chinese consider enemies to include many faces of Chinese government officials. Poison the data so that the entire enterprise is unreliable.
Old-school hacking is also necessary to understand Chinese strategy. A trove of leaked documents earlier this year revealed a great deal of information about the extent and methodology of Chinese surveillance, and this sort of thing needs to happen more often. Here is where Israel can contribute a great deal, although it requires more expertise on Chinese thinking than Israel probably has now.
Similarly, injecting our own backdoors into Chinese software libraries and source code now can pay dividends later. Supply-chain attacks are effective in both the physical and cyber spaces.
Like it or not, AI is the future. We need to take advantage of current Western AI superiority to come up with better cyberdefenses and better cyber-offenses to not only stay ahead of the Chinese threat but also to hamper it.
The Chinese already invest a lot of resources into infiltrating Western technology companies and universities to steal the latest innovations. Not only does the West need to do a better job monitoring the spies, but it should be doing the same to Chinese military and commercial AI researchers, both for intelligence and potentially for sabotage. In retrospect, the US government actions against Israel's Pegasus spyware were shortsighted - because technology like that, which can turn phones into spy devices, are what is needed to stay ahead of the enemy.
Of course, major companies like Microsoft and Amazon must do a better job in defending themselves from not only today's attacks but from novel attacks coming that are created or improved by AI. World governments are depending on these companies to protect their own data, and as we've seen just this past week, even the most advanced software companies can screw up and cause massive outages. Here is another area Israel, as a cybersecurity leader, can be used effectively. There is a reason Google is about to pay a record amount for a relatively new Israeli cloud cybersecurity company.
The cognitive warfare component is not as science-fictional as the article implies. We already have years of data on how nation-states have manipulated - or failed to manipulate - public and political leaders' opinions. An AI crafted social media campaign is certainly possible, and we are not far from an AI-created campaigns of fomenting antisemitism or anti-American feelings in Europe and Africa, for example. Perhaps it is already happening. The West needs to do the same to increase discontent within China (and Iran, Russia and North Korea), and AI can brainstorm ways to do that.
Western AI needs to be ahead of Chinese AI even in doing evil things like DNA hacking (or building an undetectable suitcase nuclear bomb.) We will need AI to come up with ways to defend against these, but we will also need AI to even identify them before they happen. Laws that protect us from the negative repercussions of AI can also hamper our ability to combat it, and we need to carve out carefully created exceptions to not subvert human rights while being able to anticipate and if necessary attack the enemy in ways we cannot even imagine. It is scary, but not as scary as the alternative of unchecked Chinese dominance in the field.
The war is happening now. We need to treat it as such.
|
Or order from your favorite bookseller, using ISBN 9798985708424. Read all about it here! |
|
