Threat posed by super-intelligent AI is down to egotism of Donald Trump and others – Martyn McLaughlin
The closest thing the world has to an AI safety inspector is in the UK, giving this country an opportunity to provide much-needed leadership
The well-publicised warnings by artificial intelligence researchers about the technology’s existential threat to humanity have intensified calls for global regulation in recent days. The potential peril of AI is, in the eyes of many, comparable to the dangers of the nuclear arms race.
Indeed, some experts have seized on the analogy to advocate a modern-day multilateral global treaty for greater AI safety, similar to the treaty on non-proliferation of nuclear weapons, which curbed the worldwide spread of nuclear-armed nations.
It is a well-intentioned but ultimately misguided comparison, not least because it ignores the fact that it took a staggering 23 years after the creation of the first nuclear weapon for the treaty to be ratified; even if the omens raised by AI staff prove premature, no one can afford to wait two decades for change.
READ MORE: AI is a super-intelligent psychopath that could kill us all. Here's what Scotland needs to do
A new industrial revolution
Such a vision does, however, recognise the importance of achieving worldwide consensus. No matter what guardrails or rules are proposed, they will be worthless without a robust regulatory framework that spans the globe.
Quite how that is achieved is one of the defining challenges of our age. The global agreements around AI that have been ratified to date, such as the 2023 Bletchley Declaration, signalled progress, but are ultimately little more than statements of intent. With Goldman Sachs estimating that cumulative global AI spending through the end of the decade could reach as much as five per cent of global gross domestic product, there is a race to achieve dominance in what has been described as the dawn of a new industrial revolution.
Unfortunately for humanity, the pacesetters are Donald Trump and Xi Jinping, with the leaders of the world’s two largest economies engaged in a zero-sum race for dominance. The two men are set to meet in Washington later this month, with AI safety talks reportedly on the agenda. Apologists for the US president, as well as those who understand his motivations, have suggested that a potential deal could secure his long-lusted-after Nobel Peace Prize.
Egos, economics and instransigence
It is a theory whose repugnance is matched only by its absurdity. Even if Washington and Beijing were to strike a pact, a bilateral agreement would do nothing to address the dangers posed by any number of rogue states or nefarious actors around the world.
In any case, Mr Trump has made clear his goal of ensuring the US becomes the world’s pre-eminent AI superpower, while blithely dismissing safety concerns as a “hoax”. All the while, efforts at ushering in regulation have been undermined by Republican dominance and lobbying campaigns.
This perhaps gets to the crux of the issue – the potential threat posed by AI stems not from the legacy of super-intelligent systems capable of recursive self-improvement, but egotism, economics, and geopolitical intransigence.
In such circumstances, it is easy to be despondent. And yet, perhaps there is a role for Britain to play in all this, having spearheaded the Bletchley agreement, which saw 28 governments, including the US and China, agree that AI poses a potentially catastrophic risk to mankind.
The internationally respected, UK government-backed AI Security Institute has rightly been described as the closest thing the world has to an AI safety inspector. If its expertise can be married to some old-fashioned diplomatic clout, there is a chance that consensus could prevail.
Comments
Want to join the conversation? Please or to comment on this article.