Vladimir Putin, of all people, was the first world leader to grasp what a big deal artificial intelligence was for geopolitics. “Whoever becomes the leader,” he warned a group of schoolchildren more than five years ago, “will become the ruler of the world.” What, back then, seemed even to cutting-edge researchers like hyperbole, feels to many like imminent reality in the uncanny age of America’s ChatGPT and China’s Ernie Bot. This is because the geopolitics of great power competition is about to be reshaped by the AI revolution. There will be no research pause, no “pause” to assess its risks that most of the world’s leading researchers are calling for. We are on the brink of an AI arms race with implications for us all.
This is exactly what Geoffrey Hinton, the AI pioneer who recently resigned from Google, wants to prevent. Dubbed “the godfather of AI”, the British-Canadian researcher believes these models are only a few years away from outstriping their human creators, will soon be exceptionally good at manipulating them, and if they were to be programmed with the ability to set their own goals would quickly realise that seizing as much control over human society as possible was a rational way to achieve it.
It is important to note that AI sceptics think this is overdramatic. They identify Hinton as making a leap of faith that a machine would even want to set its own goals if it was enabled, and part of a historic pattern of technological creators and inventors believing their inventions have magical powers.
But even the sceptics are worried that AI in the wrong hands could give immense powers to bad actors. Invent new diseases? Outline how to print them? The prompts are easy to imagine and totally possible. This is why nobody disagrees with Hinton’s current call for “nuclear-style” agreements between China and the US to encode safety structures, from an “off switch” to some kind of “do no harm” circuit, into every form of artificial intelligence available.
This is the geopolitics of AI that we should be getting. Instead, the geopolitics of AI we are about to get is the exact opposite. Both the US and China are deeply frightened by Putin’s prediction — and by what catastrophic implications there could be if they fall behind. Washington is worried that in terms of state funding Beijing has already overtaken them in terms of spending. Beijing, in turn, is worried that in terms of technology prowess Chat-GPT and American AIs seem superior to Baidu’s Ernie Bot. Instead of looking for an agreement to work collaboratively for AI safety the US is seeking to kneecap an increasingly hostile China’s AI lead and the FBI warns China is engaging in rampant high-technology espionage. The “chips war”, where Washington placed export controls on critical computing components, are best understood as AI wars: without access to such chips, the bet in Washington is that they can slow China’s technological advances and establish a dominant lead.
Artificial intelligence is best understood as like electricity. It is applicable to almost everything. Militaries and intelligence services understand this better than most, which is why the Pentagon and the CIA are investing significant resources into finding its applicability. High on their concerns are AIs such as Midjourney which can supercharge propaganda and disinformation or guide armies in the field. The AI arms race is well and truly here.
Sceptics worry that the mindset now gripping Washington will only further meld Silicon Valley into the national security state and hype of all kinds will be pushed by developers to get the biggest contracts.
What is happening is the exact opposite of the “research pause” that leading researchers want. Instead, each state, researchers warn, may throw untold resources into competitive god-like AIs.
It is a damning testament to our age that one of humankind’s greatest discoveries is being met with such fear by its own creators like Hinton. If humanity was politically organised in a way that was genuinely democratic and collaborative — not pitted against each other by superpowers and corporations — the AI research “pause,” safety embeds and regulations it needed to unleash wonders for medicine, science and human productivity. Instead, it seems set to unleash wave after wave of deeply disruptive jobs losses, weapons, propaganda and human struggle.
