Andy Burnham has arrived at the United Nations with an admirably simple solution to one of humanity’s most complicated technological problems: global cooperation.
Splendid.
The Prime Minister wants Britain to become an “honest broker” bringing major powers together to establish common principles for artificial intelligence. He argues that Britain’s relationships with the United States, Europe, China, India and Brazil make it well placed to convene those discussions, with AI at the centre of Britain’s G20 presidency in 2027.
There is, however, a diplomatic wrinkle.
Nothing Says “Come to the Table” Quite Like Hitting Someone With the Table
In the same speech about bringing a divided world together, Burnham launched one of his sharpest attacks on Russia.
He accused Moscow of spreading disinformation, cited alleged Russian interference in the 2019 general election and claimed the Kremlin spends around £1.3 billion annually manipulating information. He also announced a National Centre for Information Defence to “detect, attribute and disrupt” hostile state information attacks.
Those are serious allegations, and the British government is entitled to respond to what it considers hostile activity. But the timing creates an obvious contradiction.
Burnham’s proposal depends on countries that distrust one another agreeing to restrain themselves in a technology race worth potentially trillions and affecting defence, intelligence, industry and national power.
The message is essentially:
“Humanity must put aside its differences and establish trust.”
Then, a few paragraphs later:
“Except those people over there. We’ve had quite enough of them.”
It is less an international consensus-building exercise than organising a neighbourhood barbecue while shouting allegations over the fence.
The problem is not limited to Russia. Donald Trump has already signalled that Washington does not necessarily share Burnham’s enthusiasm for international limits on advanced AI. Sky News reported that Trump rejected a global scheme to control AI, framing the contest more bluntly: whoever leads in advanced intelligence gains the advantage.
Meanwhile, America and Britain are developing an AI defence partnership intended to protect critical infrastructure and accelerate military applications.
So the geopolitical equation is:
AI is too dangerous to become an international arms race.
Therefore Britain will establish international standards.
While entering an AI defence partnership with America.
While confronting Russia.
While hoping China joins the conversation.
Simple.
The theory behind Burnham’s proposal is not absurd. AI genuinely crosses national borders. Cybersecurity, autonomous weapons, deepfakes, disinformation and frontier models cannot be controlled neatly by regulations written in Westminster. The United Nations has also pushed for greater coordination on AI safety.
The difficulty is imagining Moscow, Washington, Beijing, London and Brussels suspending strategic competition because Britain has booked a conference room in Manchester.
Russia is unlikely to abandon its interests because Andy Burnham asks nicely. China is unlikely to surrender a technological advantage because Britain produces an excellent communique. And the United States is unlikely to slow its AI industry merely because everyone agrees that cooperation would be sensible.
This is the fundamental problem with grand international declarations: agreeing that something is dangerous is much easier than agreeing who gets to develop it first.
Nuclear weapons taught us that. Cyberwarfare taught us that. Military drones are teaching us that. Artificial intelligence is unlikely to look at thousands of years of geopolitical competition and decide that this time everyone seems trustworthy.
Suppose Burnham gets his summit. There will be motorcades, translators, security cordons, working groups, panel discussions and enough recycled conference lanyards to strangle a rhinoceros.
A declaration will emerge. Countries will promise “responsible development,” endorse “shared principles” and perhaps create a “framework.” Beneath the diplomatic stationery, however, American, Chinese, Russian and European engineers will continue trying to make their systems more capable than everyone else’s.
The real question is not whether governments can agree that AI needs safeguards. They probably can. It is whether they will accept those safeguards when complying might leave a strategic rival technologically stronger.
That is where the inspirational speeches encounter something more powerful than artificial intelligence: human self-interest.
Burnham therefore faces an awkward challenge. If Britain genuinely wants to become a diplomatic bridge between competing powers, somebody eventually has to walk across it. At the moment, he appears to be standing in the middle, waving everyone towards him while putting up a sign at the Russian end saying HOSTILE ACTORS THIS WAY.
Britain can convene the conversation, but it cannot credibly lead global AI regulation while simultaneously deepening military AI ties with Washington and publicly treating one major power as an adversary. Leadership requires trust, consistency and leverage; Britain currently has some convening power, but not enough of the other two.
So the likely outcome is not a single global rulebook, but a familiar diplomatic compromise: shared principles, flexible language and national governments carrying on as before. Britain may host the table, but it cannot honestly claim to control the menu.
🔥 Challenge: Can Burnham persuade rival powers to accept common AI rules while Britain deepens its defence-AI alliances and confronts Russia — or will global AI regulation become another declaration that everybody signs and interprets however they please? Tell us what you think on the Chameleon News blog.
🔥 Comment. Like. Share. And tell us whether global AI cooperation is serious diplomacy — or the world’s most expensive group chat.
Leave a comment