There are wars fought with artificial intelligence and there are conflicts that could be fought to ensure access to the technology needed to produce artificial intelligence. Let's start from the second type and then from the Hardware. Next we will look at software, i.e. artificial intelligence and how it can work in a war context.
Modern war is also a war of microchips. Among the first initiatives carried out by the West to put Russia in difficulty was the attempt to block the export of sophisticated microchips to Moscow. So much so that the Russians have started modifying the microchips of common washing machines for use in their tanks. The progress in this area has been such that a common washing machine chip today is significantly more intelligent than the chips used in weapons systems of the 1990s. Furthermore, many microchips made in the US are sold to China. It is not a trade that can be stopped and from there, according to many sources, they are triangulated in Russia. At the same time, however, China, which depends on US microchips in many ways, threatens Taiwan, where the world's most sophisticated microprocessor factory is located. Nobody makes and miniaturizes semiconductors like Taiwan's Semiconductor Manufacturing Company. To be clear, this company's A 14 microprocessors each contain 11.8 billion transistors. Just sixty years ago, just to give you an idea, there were 4 transistors in a single microprocessor. This makes the leap in computing power clear. Without those processors, the planet would shut down. If Taiwan fell into Chinese hands… If you want to learn more, just read the essay Chip War (Garzanti, 422 pages, €22) by international history expert Chris Miller. But let's get to the war with artificial intelligence. Obviously in the military sector the use of these intelligences, which are adaptive and capable of processing an enormous amount of data, immediately appeared interesting. Compared to normal software, they are much simpler to use and can assist human beings.
But let's start with a concrete example of what an integrated artificial intelligence system can do. In March 2022, many expected Russia to eat Ukraine in a matter of weeks. Which, among other things, seems to have been in the plans of the Russian general staff, but it didn't happen that way. What made the difference was the technological element.
The most advanced technologies, mainly American, are playing an essential role in ensuring Ukrainian survival on the battlefield, at least as regards battlefield intelligence and reconnaissance operations. «We build software that allows organizations to effectively integrate data, decisions and operations» says the website of the US technology company Palantir. The company, which declares around one and a half billion in turnover in 2021, took off in 2001, when one of the American intelligence agencies, the Central Intelligence Agency (CIA), requested its services to integrate otherwise scattered information to conduct operations of counter-terrorism. Today, thanks to Palantir, a Ukrainian soldier sits in front of a laptop and looks at highly detailed digital maps showing the battlefield, mostly obtained from commercial satellites: about forty fly over Ukraine every day. With a few clicks of the mouse the soldier sees thermal images, which show artillery fire, tanks, the disposition of enemy forces. With a few more clicks you use a targeting program to select a missile, artillery piece, or armed drone to attack the target. This is the digital war being fought in Ukraine. As cynical as it may seem, it is possible to say, in this sense, that the country is today the first practical example of how wars will be fought in this century.
Algorithmic warfare, as Palantir CEO Alex Karp called it, uses a digital model of the battlefield, allowing command centers to clear away the fog of war. What makes this system revolutionary is that it aggregates data from commercial providers. In short, artificial intelligence accelerates the OODA cycle, the observe, orient, decide and act decision-making cycle, developed by United States Air Force military strategist Col. John Boyd, who applied this concept to the phases of combat. If you play chess with a computer you will see that its OODA loop is much faster than yours. On a battlefield given the enormous variables, human intuition had the advantage. Now with these adaptive systems it may no longer be the case. We don't have to fear a terminator chasing us and shooting at us. The humanoid interface for AI is dysfunctional at the moment. What we have to fear at the moment is a computerized general. Also because as adaptive and fast as AIs are, we ourselves insert cognitive biases into them. A simulation was conducted on commercial models which can give an idea. A research team from three different US universities (Georgia Institute of Technology, Stanford University and Northeastern University) conducted a series of experiments, starting from a simple idea: inserting four different artificial intelligence models into a program developed by 'Hoover Institute simulating international crisis scenarios. In some cases, software have even resorted to the use of nuclear weapons, justifying their virtual actions with the pursuit of world peace.
Obviously the experiment is valid up to a certain point. But the fact is that these artificial intelligences have a strong assertive ability. A country at war that achieves a series of successes with an artificial intelligence will be willing to ignore its advice if at some point it seems morally unacceptable.
Or, morality aside, who tells us that the artificial intelligence that helped us yesterday won't be hacked tomorrow, to give us bad advice? We are heading into unknown territory.