The US-UK partnership on AI safety is a damp squib
Let’s break it down.
On one side of the Pond the government is heading into an uncomfortably tight election. On the other, the government is facing an electoral wipe out. Expect many crowd-pleasing, but vacuous, announcements from both. Few, however, will be as jointly coordinated as the US-UK partnership on AI safety. But it won’t do anything.
On the face of it both countries are the obvious choices to provide global leadership… actually, cough, the US is the obvious choice and the UK is it’s obvious choice of “best friend” on the issue.*
( * The UK, powerful little island that we are, has the world’s third largest tech sector. Only the US and China are bigger in this space, and Chinese-American relations won’t become cuddly anytime soon. So expect Washington to “lean on” London for support over the coming years.)
The US is, probably, the country most at risk from AI as well as the country that has the most to gain. On the downside it has a highly polarised electorate and a simmering culture war that has already exposed American democracy’s vulnerability to subversion. Few of us neutral onlookers are expecting November to January to be an entirely happy time in America.
On the upside, the US sits at the top of the techno-feudalism tree. Pretty much everything technological in the West is a downstream application of something else under US control. And that something else gets a slice of the revenue.
The UK, by contrast, has shown itself to be remarkably stable in the face of Brexit and the seemingly endless toppling of Prime Ministers. And, as hard as some on the extreme ends try to spark a US-style culture war, we just don’t really care who uses which bathrooms.
We are, however, a high-tech and service-based economy, and very weak in areas that won’t be directly impacted by the recent advances in AI. At the literal grass roots, England’s fabulously productive agricultural land could only sustain us if we revert to seasonal consumption patterns. As for manufacturing, steel, energy and so on… forget it.
But a high-tech and service-based economy does, in gross terms, stand to gain massively from AI. The UK has a rich data landscape ready to be leveraged and a highly skilled workforce ready roll up its sleeves and…
… there’s the rub. What happens to that workforce if companies and public bodies adopt a cost-saving mindset to AI instead of a productivity gain mindset?
Everybody in tech knows the answer — mass layoffs. They can happen in any developed country, but the UK is looking particularly vulnerable as it wallows in post-pandemic stagflation.
So, what can two embattled governments offer to their jittery people that looks like leadership?
A wishy-washy memorandum of understanding, of course! One in which they “intend” (that’s the word they use) “to work closely” and “perform at least one joint testing exercise on a publicly accessible model” and “intend to explore personnel exchanges” and “share information … in accordance with national laws and regulations, and contracts” (… hmm).
“But, an MoU is a solid first step towards working together, right? I mean, doesn’t the act of just making one set the stage for further progress?” Says one reader who I’ve made up in my head.
Well, dear imaginary reader, sort of. And, if this was between two governments that don’t ordinarily play nicely together, it would be more worthwhile news. But the US and UK are so close that one supplies the other with nuclear capable missiles (which are still a far more efficient means of global destruction than super-powered chatbots).
Announcing that they “intend” to “explore” things around a policy area isn’t much to write home about.
The question we ought to be asking is what happens to this partnership next year when the UK has a Labour government (and nobody seriously expects otherwise) and the US has either a power-boosted Biden or Trump?
New Prime Minister, Keir Starmer, will almost certainly want to build a centrepiece of UK-US relations around an area that doesn’t involve one of the wars going on (as a people, we really are done with all that). But an invigorated President will almost certainly want to focus on assuming the role of peacemaker.
Will the man in the White House want to get bogged down with the thorny issue of restraining AI companies — who are making more money for the US than they are for anyone else — when there are far bigger political wins north of the Black Sea and East of the Mediterranean?
Not likely.
Tell me I’m wrong.
