Countries are being pushed into a future of AI rivalry. Of the US versus China. But there’s an alternative foreign policy for us all to pursue – built on cooperation
“If you work in this space, sooner or later you are going to have to choose.”
Ash Fontana is an AI investor and author with that rare quality in a complicated world – of being able to cut to the chase. If you are at the cutting edge of science and technology, you are going to have to takes side, either making your home in the US marketplace or China’s ecosystem; you can’t have both.
Masayoshi Son, the founder of Softbank and Vision Fund investor, now says that all his investments in the coming decade are focused on only one thing: artificial intelligence. AI, in a word he says, is about the power of prediction. And he too sees the world cleaved in two: the US accounts for about 45 per cent of the AI innovation that matters, China 40 per cent, and the rest of the world just 15 per cent.
Add into this mix – as Tom Hurd, who ran the UK’s Office for Security and Counter-Terrorism – science and technology as the new fourth chapter to global power politics, alongside diplomacy, economics and the military. And, taken together, the Clash of Civilisations has given way to the Cold War has given way to the Intelligence Match.
I’m James Harding, editor and co-founder of Tortoise, and if it sounds like I’ve been hobnobbing, I have. Good, old-fashioned asking people questions, learning more than you knew. Yesterday, we held our first Responsible AI Forum at Waddesdon in partnership with the Rothschild Foundation; there were 100-odd people from around the world online; there were 30 people, given the rules, in the room; and there was lunch.
And over this first working lunch in a good long time, we ended up talking about the difficult issue of taking sides. Not which side to take. It’s not such a hard one: US and its principles of liberty, or China, ethnic oppression in Xinjiang, suffocation of freedom in Hong Kong, and Xi Jinping’s ever tighter command and control?
But the risk of having to take sides: the danger that the co-option of civilian businesses into this global rivalry reduces the human ties between East and West and ratchets up a contest between two value systems into misunderstandings and a disconnect that more easily spills into a conflict between two AI empires.
Choosing sides is not just a problem for the brainiacs; it’s an increasingly everyday problem for all of us. For example, the AI scientist I was sitting next to was in the middle of a wrangle within his company about opening in China. His Chinese colleagues were offended that they were considered a security risk. His US colleagues thought a Beijing office was a business risk, i.e. clients working in the US would shun them.
Eric Schmidt, the Google founder, has been a step ahead in this argument for some time. He warned of the “splinternet,” the rift in operating systems between China and the US, and he’s predicted the AI capability race will have a new leader – i.e. China edging past the US – within not five years, but “a few years”.
Next week, we’ll publish a considered think on what we learned about Responsible AI, amazing AI (fusion, accelerating net zero, protein folding, hard maths problems), and chilling AI (unaccountable monopolies, cyber dark nights, crypto money laundering).
But to coincide with the AI Forum, we published the update of the Tortoise AI Index. It’s complicated to put together, but easy to read: and it shows, as Masa Son sees it too, the rest of the world lagging further behind the US and China.
The implications of all this for the UK – and every other country, really – are profound. You can’t help fearing that this next wave of technology is driving us apart.
But surely – and a sentence starting like that is usually a sign that this is not a fully formed thought or it’s a dash of naive idealism – there is a foreign policy vision that plots an alternative path to this AI rivalry.
It comes in two parts. One, the prohibition of certain kinds of AI can prevent division over all kinds of AI. Tom Hurd’s argument at the AI Forum, for example, was that we need to agree some things now: i) there’s always a human hand on the nuclear button; ii) that there’s no use of AI in hot zones, i.e. war zones; iii) a global alliance prohibiting the use of AI in bioterrorism. If certain kinds of AI are outlawed, then it will leave open the possibility of more global collaboration – yes, China and the US included – in areas of everyday life such as health, education, logistics, transport and energy that can be transformed by machine learning, data and AI.
The second part is an AI internationalism. Demis Hassabis, the founder of DeepMind, suggested the possibility of a World Institute of AI, resourced both to drive research and innovation as well as framing global regulation. For countries like the UK, ranking third in the Tortoise AI Index and with the pack closing in behind it, the pack is not the threat, but the opportunity.
We would surely be better off if we don’t have to choose, if we don’t divide the world into two rival campuses of intelligence. But we’re driving that way. And the risk of the pandemic in the early hours of the AI age is that mutual suspicions between countries, the renewed mantra of national resilience, and the US-China rivalry all conspire to prevent a new kind of globalism. One that’s unfashionable and more needed than ever. One that, if we don’t orchestrate it, will leave us in a world where we can’t take back control – because we don’t have it at all.
Europe falls behind in the AI race
A weak startup ecosystem is holding the EU back from unleashing its full potential in the AI space, while the US and China still reign supreme, according to the latest release of the Tortoise Global AI Index