Artificial intelligence (AI) seems to have developed a life of its own. We may come to regret it – Susan Dalgety

Dr Geoffrey Hinton, the ‘godfather of AI’, said some aspects of artificial intelligence were ‘quite scary’

Edinburgh is famous for being the birthplace of the Scottish Enlightenment – a revolutionary new way of thinking that helped create the world as we know it. But more than 200 years after David Hume and Adam Smith, an Edinburgh student was quietly sowing the seeds for a revolution that threatens to change humanity.

In 1978, Geoffrey Hinton was awarded a doctorate in artificial intelligence (AI) by the University of Edinburgh. While the rest of us were still using dial-up phones and choosing between three TV channels, young Geoffrey was grappling with much bigger questions than whether to watch the Kenny Everett Video Show or Dallas. He wanted to understand how the human brain thinks and how to apply that to machines which can learn. Deep learning, as it is known in the trade.

Fast forward to last week. Dr Hinton, now known as the ‘Godfather’ of artificial intelligence, hit the headlines when he stood down from his role at Google, saying that he regretted his lifetime’s work. His decades of research have led to current AI systems like ChatGPT, which have made deep-learning tools accessible to everyone with a smartphone. But, according to Dr Hinton, apps like Al Smith and Chat On are not just simple tools that help you write jokes, draft difficult emails or interpret your dreams.

In a statement to the New York Times and in follow-up interviews, he declared that some aspects of AI chatbots were "quite scary". "Right now, they're not more intelligent than us, as far as I can tell. But I think they soon may be,” he said, and he warned that there are “bad actors” who would use AI for “bad things”. He’s not alone in his fear of machines. Elon Musk, who has made billions from tech, was one of the 1,000 signatories of a recent letter which said that humanity faced "profound risks" from AI and called on its development to be paused immediately.

At the moment, apps like AI Smith provide little more than writing prompts and the occasional giggle. Curious, I asked it to write me a joke about Nicola Sturgeon. “I don’t have any personal anecdotes,” it replied. “But here’s a joke I have written.” "Do you know why Nicola Sturgeon is always so serious? Because she's always swimming upstream against the current political system! And let's be real, with all the fishy things happening in politics lately, who wouldn't be serious?"

I am not sure that the Stand comedy club would offer me an open mic spot with material like that, notwithstanding my feminist views – wink, wink – but having spent a morning playing around with ChatGPT, I can see why Musk and Dr Hinton are worried. The UK Government has just published its proposals for a “light touch” regulatory framework for AI, but I think that is too late.

The work that Dr Hinton started nearly 50 years ago in Edinburgh has developed a life of its own. Human beings love progress, and we won’t stop our pursuit of deep learning until the machines are smarter than even David Hume or Geoffrey Hinton. And what then?