IA: The owner of intelligence | Science of the week

Few technologies have been developed without the military having an eye on them. Archimedes himself, says the legend, used his studies on the reflection of sunlight to scorch with a combination of mirrors the fleet of the Roman general Marco Claudio Marcelo, and with this he earned a certain thrust in the heart by a soldier pissed off. This is what happens when a scientist gets into politics, who comes out feet first. The monetary cornucopia that financed particle physics in the second half of the 20th century was the direct consequence of the Manhattan project that, rounding it up a bit, solved the Second War to the beast. Virology will soon experience a phase of splendor, for reasons that will seem obvious to a military man. GPS was a Pentagon military tool until Bill Clinton decided to donate it to the world free of charge. Although only in exchange for financing more advanced versions of the GPS for the military.

The same we are living with Artificial Intelligence (AI), the set of mathematical techniques that has been generating amazing prodigies for a decade. Machines that see and respond to human calls – often after listening to Vivaldi’s four stations in hands-free, but that’s not the robot’s fault, but the rat that bought it – artificial legs and arms that respond to the mere thought of paralyzed people, drones and martecopters like the one that flies on Mars these days, adaptive systems that not only beat human champions of chess, go or poker, but also discover in a matter of hours strategies that had not occurred to anyone who is made of meat for centuries of culture and excellence. Humiliating? Undoubtedly. Of military interest? Of course.

The computer science elite has been clamoring for an international treaty to regulate the warlike use of AI for 15 or 20 years, in the style of nuclear non-proliferation pacts

The most advanced country in AI is still the United States, even if China is determined to eat the ground, and the liberal, almost libertarian tradition of the American giant in terms of industrial policy is a stumbling block for the initiatives to analyze the ethics of that powerful technology. What the Silicon Valley moguls do with the messages they transmit on their networks and the data they hoard in their bases matters little to Washington as long as they bring wealth to the country. But the computer science elite has been clamoring for 15 or 20 years for an international treaty to regulate the warlike use of AI, in the style of nuclear non-proliferation pacts. They know very well what they are talking about, because the Pentagon recruits them from among its hosts.

Europe has a chance there. Not only because of the war side of the matter, but also because of the daily abuses. AI algorithms are already used to select employees, for example, and we know they are biased by gender and race, since they have learned by reading our texts, which they are too. If Europe regulates, half the world will follow.

You can follow MATTER in Facebook, Twitter me Instagram, or sign up here to receive our weekly newsletter.