AI risks are something humanity can manage – let’s get working on how it can save us from ourselves

It seems a day doesn’t go by without another thought leader warning of the clear danger artificial intelligence poses to civilisation as we know it.

While they’re not necessarily wrong, I’d argue the risk of AI is something humanity can manage.

But handwringing over digital dystopia is distracting us from a very real existential threat that we’re not doing well at addressing at all – the increasingly rapid collapse of the environment which sustains us.

This isn’t to say we shouldn’t be very worried about AI.

Last month, hundreds of experts, including some of the biggest names in AI research, endorsed a statement from the Centre for AI Safety in San Francisco saying: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”

If bad actors give large AI models the necessary instructions, or good actors lose control of their creations, an AI or a group of them working together could do enormous harm.

And not because they have “human-like” intelligence but, as former Google AI “godfather” Geoffery Hinton warns, precisely because they don’t.

But while “extinction” due to AI might be theoretically possible, and it’s right to start thinking seriously about it, it’s a big word to be throwing about when a far more plausible extinction event is accelerating right in front of us.

‘Alarming’ temperatures

In the same week the latest AI doomsday proclamation was issued, a very different group of scientists started talking about what was going on in the North Atlantic.

Sea surface temperatures there are rising to a level and at a speed that’s surprised even the most cool-headed climate scientists.

The reasons for the sudden surge in ocean temperatures are complex but still “alarming” says Dr Marilena Oltmanns, a marine physicist at the National Oceanography Centre in Southampton.

“Ultimately, heat is energy,” she says. “In a warming climate, more energy is available to drive the climate system.”

That energy is what allows the atmosphere to gain moisture – increasing the intensity of storms.

Anyone trying to get home in southern England’s thunderstorms this week didn’t need an AI simulation to tell you what that feels like.

International co-operation essential

AI needs regulation, it needs international consensus.

Its potential is so huge and powerful it won’t be easy to persuade countries leading its development like the US and China to come to the table to agree on some ground rules. But history suggests that that isn’t impossible.

During the Cold War, the US and the Soviet Union agreed on a way to reduce the risk of the existential crisis of the time – nuclear annihilation.

While that threat is greater now than at any time since the Cold War, there is still some confidence in the rules to prevent it happening.

In the late 1990s, most nations adopted moratoria on the biotechnology of cloning to prevent anyone from trying to create a human life. And even in countries where there aren’t laws in place to prevent this, it’s not acceptable.

Chinese authorities reacted harshly to a doctor there who claimed to have manipulated human embryos.

Humans are brilliant at creating new and terrifyingly powerful tools like nuclear bombs and biotechnology.

AI is one such tool. But we have demonstrated an ability to regulate them.

Read more:
AI could be used to provide ‘personalised learning’ to schoolchildren, Rishi Sunak says
‘Regulate it before we’re all finished’: Musicians react to AI songs flooding the internet

A sacrificial environment

Perhaps this is an overly optimistic view – but it’s certainly more than you can say about our ability to reign in another aspect of human nature – making profits at the expense of the environment.

Despite more than 20 years of international negotiation to address climate change, carbon emissions are continuing to rise.

The environment (for example, our forests and oceans) on which all our economic output is based, continues to be depleted faster than it can replace itself.

And here’s where AI could come to our salvation.

It cannot solve all of humanity’s problems, but giving chemists powerful AIs could produce new catalysts hoovering carbon out of the atmosphere.

Physicists are already using it to accelerate the development of nuclear fusion providing limitless power without warming the planet.

Researchers at tech giant Nvidia are attempting to tap into vast computing power to create a “digital twin” of the planet that could allow artificial intelligence to guide us through some of the limited options we have left to avert an ecological catastrophe.

Human intelligence evolved over millions of years and on the current trend, by the middle of this century, the planet will be warmer than it ever was during that long evolutionary process.

Let’s stop worrying about how artificial intelligence will replace us, and instead get working on how to use it to save us from ourselves.