Let’s worry about climate change, not the ‘existential threat’ of AI
F From warnings of plague, Biblical floods, pestilence, fire and brimstone, human beings have always had a particular propensity for catastrophising, almost regardless of how material the supposed threat really is.
End of days was a core belief for many of the millenarian cults that thrived in the Middle Ages – not without cause back then, as it happens. The Black Death alone is reckoned to have wiped out roughly a third of Europe’s population.
More recently, we’ve worried about nuclear annihilation, a meteorite strike such as the one that killed the dinosaurs, Aids, pandemics and of course our old friend climate change.
All of them have been the subject of endless survivalist imaginings, warnings, alarmist forecasts, and dystopian movies.
And in virtually all cases, the risks have turned out to be grossly exaggerated, even if the jury is obviously still out on climate change. Since Putin’s invasion of Ukraine, the nuclear threat has also come racing back.
Even so, mass catastrophe, though much predicted, is in fact a relatively rare occurrence, and despite some near misses, there has been nothing throughout human history even remotely close to an extinction event.
Remember the millennium bug? This admittedly would never have finished off the human race, but an estimated half a trillion dollars in today’s money was spent preemptively updating computer systems amid warnings that it would bring the world economy to its knees if nothing was done. A far more cost-effective approach would have been simply to deal with failures as they arose, which in the end were minimal in any case.
Many of the measures deemed necessary to fight the recent pandemic similarly look like an extreme overreaction. In the event, the case mortality rate, though significantly higher than ordinary flu, scarcely seemed to warrant the myriad other harms that lockdown inflicted on the economy and public health.
So it is with a degree of scepticism that I’ve been following the latest outpouring of warnings about the threats posed by artificial intelligence (AI). Many of them come from those working at the cutting edge of these technologies, and should therefore not be lightly dismissed.
Last week’s news that an AIcontrolled US military drone “killed” its operator in a simulation to prevent it from interfering with its mission makes for particularly disturbing reading, even if such mishaps are not exactly new to military operations.
It is as if the Skynet machines of the Terminator movies are not that far off.
The possibly destructive implications for jobs are one thing. We also need to worry about AI’s potential for industrial levels of data-scraping, breach of copyright, invasion of privacy, online harm and misinformation. Fake news and invented scandals already abound; in an age of unbridled AI, it may be impossible to tell what’s real from what isn’t.
Yet let’s meet these challenges as they arise, and not attempt to preempt them. Many of the supposed harms are at this stage no more than speculation. To be stifling the industry at birth because of its potential for abuse is to deny the economy a transformative technology that offers boundless opportunity for advancement. There will be crises, there will be scandals, but that’s true of all economic development.
One thing is for sure; authoritarian regimes such as China will be suffering no such qualms. If a potentially hostile power is arming itself to the hilt with the technologies of the future, you have no option but to join the race.
Whatever one’s views about the origins of climate change, we do at least know that global warming is a real and present danger. We don’t yet know that about AI.
For real-life evidence of the already highly destructive nature of climate change, you don’t need to be guided by the counterproductive bleatings of Greta Thunberg, Extinction Rebellion and Just Stop Oil. Ask the insurance industry instead. The average payout in claims over the first 10 years of the century was in the order of $50bn (£40bn) per annum. Since then, it has doubled to $100bn, and in 2022, it was an all time record of $132bn. Part of the explanation is inflation, together with growing instances of shoddy workmanship in construction. But the overwhelming cause is climate change. Extreme weather events have grown steadily more frequent and destructive.
The trend is undeniable, and if maintained will soon render large parts of the world uninsurable against wildfires, floods and hurricanes, if not outright uninhabitable. What we also know is that globally, emissions are still going up, not down, so it is highly likely that these trends will persist, and possibly accelerate.
The threat posed by climate change is in other words a good deal more real than that of AI, which ironically could soon be offering solutions by helping to devise the cost-effective technologies needed to suck vast quantities of carbon back out of the atmosphere and bury them anew beneath the North Sea and other suitable depositories.
Turn the argument around, then, and AI should be seen not as an existential threat but as part of our salvation.
Interestingly, some of those who warn in apocalyptic terms about AI are the very same corporations that are scrambling to find ways of exploiting it. Somebody stop us, cry the tech giants, or you’ll all be sorry.
Depressingly, the disruptive yearnings of corporate youth have quickly given way to the defensiveness of middle age and established industrial success. Big tech’s faux warnings should be taken with a pinch of salt, for incumbent players have a vested interest in barriers to entry. Oppressive levels of regulation make for some of the biggest. For large companies with dominant market positions, regulatory overkill is manageable; costly compliance comes with the territory. But for new entrants it can be a killer.
For the UK, success in AI is particularly important. As things stand, Britain is third in the world for innovation in these fields, after the US and China. Across multiple other industries, Britain faces decline and irrelevance. But in the fast-growing creative and tech sectors, the UK thrives. New industries, in other words, to fill the void being left by the City, which may again be on the wane, having been substantially deprived of its once lucrative European markets.
Yet it also requires the right regulatory and tax environment.
If ever there was a Brexit opportunity, this is it. By offering a flexible, nimbler and less heavy-handed approach to regulation than that envisaged by the European Union, Britain can indeed make itself a magnet for AI investment. Encouragingly, the Government’s recent white paper on AI seems to strike about the right balance, at least in terms of its intent. But there’s many a slip, and much else the country has to get right – from enterprise-friendly tax to infrastructure and public services that actually work – before it can be certain of such a future.