That feeling came again, less than ten years later, when I first spoke to Alexa; hearing a machine answer my questions about the time and weather in a human-like voice felt like the height of techno-futurism. The distance between those moments of dizzying technological progress keeps getting shorter.

ChatGPT, launched less than five years after my first Alexa interaction, makes the “virtual assistant” seem as sophisticated as the speaking clock. ChatGPT, for those who have not yet tried it, can not only tell you the time and the weather. It can write remarkable modernist poetry about the sun in Glasgow – “in Glasgow's grey pocket tucks/ a shard of warm /glinting (a kiss for dour skies)” – and it can suggest what username Alexander Graham Bell might have used in Scotsman.com’s comments section (TalkToTheBell or CallMeAlexander).

I have written before about the struggle our political institutions face in keeping up with the (increasingly breakneck) speed of technological development. With images already circulating of AI chatbots benignly talking a person through the steps of making napalm and synthesising sarin gas, or helpfully fixing the spelling and tone of a phishing email to make it sound like a native speaker, these new technologies pose a challenge like no other.

What can governments do to keep citizens safe in the face of these threats? In his new book, Why Politics Fails, Ben Ansell argues that successful solutions to long-term challenges like these are found in institutions which allow citizens, organisation and governments to build trust in one other and provide them with opportunities to work together to peacefully negotiate their inevitable differences.

It was with this in mind that I read a recent column in The Times which argued that Australia, Canada, New Zealand, Britain and the US should form a new alliance to lead the development of safe AI technology. Leaving aside the fact that Australia, the UK and the United States recently signed a pact to “expedite cooperation in critical technologies”, we should consider the fact that this grouping of states has historically banded together, most notably in the form of the Five Eyes, because of their relative lack of substantive differences.

As the global balance of power shifts further south towards the African continent, which will be home to the world’s largest cities by the end of this century, and further east towards China and India, it should be unavoidably clear that lasting solutions to the challenges of the coming century will not be found among the same political groupings which shaped the 20th century.

If this wasn’t obvious before the Russian invasion of Ukraine, the initial unwillingness of many states in the Global South, including Brazil, India and South Africa, to condemn Russia at the United Nations should have been a wake-up call. For too long, politicians in the developed world have paid lip service to these increasingly powerful and assertive emerging powers, continuing to shut them out of decision-making processes will only further erode the legitimacy and relevance of our global institutions.

Artificial intelligence (AI) can do be used for good or evil (Picture: Getty Images/iStockphoto)

Yet, even as Google’s chief executive calls for the development of a global framework to regulate AI like the one created for nuclear weapons, it seems highly unlikely there is a global consensus to be found on these issues. The system of multilateral global governance is now being aggressively contested by China and Russia, who seek to create a new global framework that better serves their own self-interest, and the architecture which enforces international law has been crumbling around our ears for decades.

Rather than look to familiar faces or the floor of the UN General Assembly, the UK might look to form a “Goldilocks alliance” – one that’s neither too small nor too big – to address the urgent threats that these new technologies pose. It can do so by finding partners from across the world – Mexico, Brazil, Nigeria, South Africa, Japan, South Korea – in addition to its traditional allies, and working with them to shape the development of global rules that will survive the challenges of the century to come.

This will be no small task. While the existentially destructive potential of nuclear weapons was evident from the start – the director of the Manhattan Project watched the first nuclear bomb explode and called himself “Death, the destroyer of worlds” – the same cannot be said for technologies like artificial intelligence. It is not fully clear what harm, or good, they may yet bring us.

Because of this, the UK has argued at the UN that binding global regulation on cyber and digital issues would represent “the partial and premature imposition of an approach to a domain that is currently too immature to support it”. As the UK’s language shows, powerful states are unlikely to submit to any kind of legal framework which would tie their hands when it comes to using as-yet-unknown technologies which may offer them the upper hand in an age of strategic geopolitical competition and weaponised interdependence.

But consider how this technology, for good or ill, will have developed and come to be used during the next pandemic, future elections, a critical national incident or even a hot war. Finding some form of consensus, norms and rules will be difficult, but the alternative – no rules, standards or norms – is almost unthinkable.

Governments cannot continue to play catch-up as world-changing AI applications are rolled out to the public. They must grasp the issue with the urgency and the global perspective it requires. Given the exponentially increasing speed at which these technologies are developing, there is increasingly little time to waste.

Stewart McDonald is SNP MP for Glasgow South

As artificial intelligence (AI) helps terrorists and online scammers, UK should look to form a new global 'Goldilocks alliance' – Stewart McDonald

Comments