How worried should we be about AI? Absolutely petrified, according to a new documentary called Chasing Utopia. Billed as the Inconvenient Truth of the digital age, it’s a shop window for the alarmist predictions of Mo Gawdat, a former chief business officer at Google X who’s reinvented himself as the Al Gore of AI. The filmmakers follow him on a global speaking tour, delivering the same TED talk to interchangeable groups of credulous young people. ‘Beware!’ is his message. If we don’t rapidly change course, we’ll soon be at the mercy of a superior alien intelligence that sees humans in much the same way we view chimpanzees.
I don’t buy it – partly because I think we’re a long way from artificial general intelligence, let alone artificial superintelligence. Gawdat talks as if a malevolent, HAL-like entity is slouching towards Bethlehem, but plenty of other people in the sector reckon that AI systems based on large language models (LLMs) are never going to evolve beyond being very, very good at domain-specific tasks. Yes, coders, lawyers and accountants should be worried. The rest of us? Not so much.
The new Jeremiahs have emerged just as the bottom fell out of the market for climate alarmism
The main reason I’m sceptical, though, is that catastrophising about AI seems like yet another elite boondoggle. If you’re a nerdy founder of a tech company, an academic computer scientist or a failed politician, scaremongering about ‘existential risks’ is a surefire way to reinvent yourself as a public intellectual. Indeed, describing Chasing Utopia as the Inconvenient Truth of AI rather gives the game away. It is surely no coincidence that a new cast of Jeremiahs have emerged – the ApocalyptAI – just as the bottom has fallen out of the market for climate alarmism. Even Greta Thunberg has abandoned green activism for Palestine. Now the climate Cassandras have moved on, men like Gawdat – and Geoffrey Hinton, billed in the film as the godfather of deep learning – have popped up to fill the gap.
And gap there is. Not only does the educated elite have an inexhaustible appetite for secular millenarianism, but they particularly like doom-mongers who identify profit-chasing capitalists as the greatest threat to humanity. For the environmentalists, the font of all evil was Big Oil; for the ApocalyptAI, it’s Big Tech. Palantir is the new Union Carbide; ‘tech bros’ like Peter Thiel and Elon Musk are the new Koch brothers.
Why do high-status liberals lap up these attacks with such enthusiasm? Because the way to mitigate the risks we’re being warned about is more regulation, more intergovernmental committees, more academic institutes, more thinktanks, more seats on boards for ‘ethicists’ – more jobs for people like them, in other words. Only the expert class can be trusted to oversee the rollout of AI, even if their track record on managing other ‘existential risks’– the financial crisis of 2007-08, the explosion of radical Islam, the Covid pandemic – is less than stellar.
But let’s not get distracted by tiresome details. The argument for worrying about these threats is always emotional, never rational – and Chasing Utopia is a case in point. Yes, we get a few facts about autonomous drones going rogue and LLMs trying to deceive their trainers, but they’re submerged by visual images straight out of a cheesy horror film, such as alien, long-fingered hands pushing outwards from inside amniotic sacs.
There are other reasons to be dubious about these prognostications of doom. Every new technology has been accompanied by elite panic, from the printing press to television, yet the world still turns. You only have to look at the public intellectuals’ climate change forecasts – Al Gore predicted Arctic summer ice would be gone by 2014 – to realise just how unreliable these soothsayers are. We’re supposed to accept the premise that free markets alone can’t prevent Armageddon, but AI companies have an incentive to prioritise safety without the need for more regulation. If a rogue chatbot escapes from their R&D labs, it will destroy their businesses. Then there’s the standard objection to more red tape – it favours well-resourced incumbents who can afford compliance departments over scrappy start-ups. The way to keep the AI sector healthy is to encourage diversity and competition, not stifle it.
But perhaps the strongest objection to prophets such as Mo Gawdat is that the rapid growth of AI and its deployment in areas like law enforcement and security do pose genuine risks – not to the future of mankind, but to civil rights. It’s those problems policy-makers should be focusing on, not the sci-fi fantasies of the ApocalyptAI. Films like Chasing Utopia are an unhelpful distraction.
Comments