How would you treat an ant? It’s right there on the pavement in front of you – you realize as your leg is already on its way down, the shadow of your foot cast over it, its fate seemingly decided. All it would take is the jerk of a muscle and the ant would be out of harm’s way. But its life is so unimportant, so expendable. It’s hard to inflate the meaning of one ant’s life to something more than it is, to something we should care about. When they pose an obstacle we kill them with utter disregard. There is currently no being out there which stands above humans in the same way, but experts in Artificial Intelligence say that there one day could be, and depending on who you ask, it’s either humanity’s Holy Grail or Pandora’s Box. Perhaps it should worry us then that the stage seems set for an A.I. arms race.
Self-proclaimed canary in the coal mine and CEO of Tesla and SpaceX Elon Musk claims that A.I. will become far more dangerous than we can currently comprehend, insisting that humanity treads carefully. He differentiates between two kinds of A.I.: general machine intelligence and case-specific applications of machine intelligence, like self-driving cars, with which he’s quite familiar. He refers to artificial general intelligence as having an “open-ended utility function”, allowing it to be applied to any problem, not just a specific task, and that it is this kind of A.I. we should be paying the most attention to.
“I am not really all that worried about the short term stuff. Narrow AI is not a species-level risk. It will result in dislocation, in lost jobs, and better weaponry and that kind of thing, but it is not a fundamental species level risk, whereas digital super intelligence is,” explained Musk.
To some, worrying about super intelligent AGI (artificial general intelligence) is like worrying about overpopulation on mars, a planet we haven’t even landed on yet. That may be true; it does not however prevent us from exploring how the high-tech showdown is shaping up, and China seems to have some notable advantages.
America and China are heavyweights in the field of A.I. Their tech-giants have access to the largest amount of data, the best talent and boast the biggest computing clouds – all crucial ingredients needed to develop A.I. services. The US has long dominated the industry, with Google, Microsoft and Amazon leading the charge, but the gap is shrinking and China is catching up. The Pentagon estimates that total American spending on artificial intelligence systems in 2017 was “a couple of billion”, while China is pushing the $12 billion mark, a number expected to grow to at least $70 billion by 2020.
This year China invested $2 billion in infrastructure projects to house hundreds of AI-companies in a dedicated A.I. development park in western Beijing. Chinese A.I. companies are highly specialized and numerous. Educational policy is geared towards producing more STEM (science, technology, engineering and mathematics) workers. And crucially, it’s compulsory for businesses and private citizens to share their data with the government. Self-learning machines need pools of data to “practice” on – and the more there is and the more detailed, the better. In the world of A.I., data is as close to gold you’re going to get, and in this regard China could be king Midas.
Predictions like these lead many to believe that China is well equipped to achieve its goal of becoming the world leader in AI by 2025. Its progress towards this goal seems largely unchecked.
The kind of A.I. in which China aims to be the leader is very narrow indeed, however. In essence, it consists of the mass implementation of AI geared for specific tasks into the wider society. The rapid adoption of AI-based solutions in China might be due to the “leapfrog effect”: Chinese citizens jumped directly to mobile-payments, never truly adopting credit cards. Chinese hospitals never had established traditional digital systems for booking patients, so they are jumping straight to smartphone apps; and so on.
Public officials are encouraged to plan in accordance with these realities, so that private A.I. companies can develop products that can actually be widely used. What they end up getting is troves of detailed data on their citizens, posing an entirely different dilemma as to what China will do with all this information. Some have argued that A.I. will play a key role in the contentious social credit system, where citizens will be rated according to how appropriately they conduct themselves in society, down to details like how carefully they cross the road. Those of you who on many occasions speak the word “Orwellian” might want to start shouting.
The use of autonomous weapons systems also relies heavily on A.I., and China has proven itself to be up to the task. By 2016, the country had demonstrated its capacity to use autonomous weapons technologies in each domain: on land, in the air and in the sea. The 2017 arms-control advocacy video Slaughterbots has its creator Stuart Russell, professor of artificial intelligence at Berkeley, sign off by saying: “We have an opportunity to prevent the future you just saw, but the window to act is closing fast”.
The experts’ warnings regarding artificial general intelligence are frightening, but so are some the potential applications of artificial narrow intelligence. China has positioned itself well to become a leader in the implementation of artificial intelligence and many see it as having an edge over the US. While it is hard to compete with the gravity and severity of Musk’s prediction about sci-fi self-extermination, the successful implementation of artificial narrow intelligence in China could be ground-breaking in many respects. It is not yet time for us to come to grips with super-intelligent machines as a factor of daily life, in China or anywhere else. Until then, it wouldn’t hurt to think about how we would want these artificial and generally intelligent machines to treat us. How much thought do you give that line of ants industriously tunneling into the cracks of the sidewalk? How would you treat an ant?
Dino Ekdal
