New warnings that humans are flirting with out-of-control artificial intelligence raise an uneasy question: Has AI surpassed climate change as one of humanity’s gravest risks?
Although the danger of the technology is hard to assess, the worst-case scenario – human extinction – makes other earthly concerns seem small.
Yes, climate change is creating dangerously hot conditions, fueling superstorms and killing crops, and it may lead to planetary upheaval. But that only matters if our civilisation is still recognisable.
“In terms of things that could be the end of humanity, I think AI is a bigger risk,” said Toby Ord, a researcher at Oxford University and author of The Precipice: Existential Risk And The Future Of Humanity.
All of this leaves climate activists competing for a finite pool of political consideration. (As diplomats and world leaders gathered at the United Nations last month to grapple with the planet’s most pressing problems, some worried that global warming could be overlooked even as it worsens.)
Leah Stokes, a professor of environmental politics at University of California, Santa Barbara, wrote last month that AI was “eating up all our attention” and argued that climate change posed the more concrete danger, citing deadly heat waves in Europe this summer and a horrific glacier collapse last month in Nepal.
“Unlike AI, climate change’s existential threat is not hypothetical. It is killing people now,” Stokes wrote. “It also feels like a lot of people have conveniently forgotten about the urgency of solving this problem.”
Humans are notoriously poor at assessing dangers, and in the case of artificial intelligence, the risks are largely speculative at this point. Even the boldest AI extinction predictions often say little about how the disaster could unfold. Would the end come in the form of an AI-engineered superbug or the launch of nuclear weapons? The destruction of human society could also unfold more slowly, with a rogue takeover of infrastructure as humans go hungry and lose access to medicine.
“There are gradual AI catastrophe scenarios,” said Seth Baum, executive director of the Global Catastrophic Risk Institute. “We continue to automate more and more over time. And we end up with something that is fully automated. And something goes wrong.”
The predictions stem from the sense that humans are creating a potentially untamable entity, one that could gain the ability to self-improve and become smarter and faster than us in every meaningful way. The solutions to thwart all this? Not yet in place.
Climate change, on the other hand, is full of knowns. The science linking fossil fuel emissions and global temperature rise is long-established and irrefutable. The solutions to help curb emissions, such as solar energy and electric vehicles, are proven and abundant. At the global level, the annual conferences aimed at averting the crisis have become a fixture on the international calendar, replete with scientific and political jargon along with diplomatic wrangling over long-term action plans.
That may make them feel more bureaucratic than existential – but they’ve proved meaningful. Global temperatures had once been projected to rise 4°C beyond preindustrial levels. Now scientists say warming is on track for 2.6°C, based on the current policies and trends.
Still, that’s well beyond the level that nations said would be tenable.
The consequences of climate change are showing up in extreme bursts such as wildfires, heat waves, floods. At the same time, it is also building relatively slowly, in ways that test the attention span. Since fully landing on the world’s radar a generation ago, the issue has sparked moments of collective unity, as well as backlash, despair and fatigue.
“Things that happen gradually, we just sort of normalize them and get used to them,” Baum said. “Ordinary people don’t want this to be the thing they’re always thinking about.”
Human-caused headaches
As threats to humanity, AI and climate change have some notable links.
In both cases, the risks stem from human-made pursuits, whether it’s the development of advanced computer “frontier models” or the burning of fossil fuels that powers the modern global economy. And in both cases giant companies (trillion-dollar tech firms or globe-spanning energy giants) hold outsize influence over the course of history.
Trump has discounted warnings about AI, saying on social media recently that they come from the “same people that said, just a short time ago, that the World would be extinguished by ‘Climate Change’”.
AI has the potential to be a tool in the fight against climate change. It could design more sophisticated warning systems for disasters, for example, or dream up innovations that ease global water shortages. Already, its systems are being used to help fortify the low-lying Maldives and regrow shoreline.
But in the short-term, at least, AI is making it more difficult to cut greenhouse gas emissions. That’s because it is increasing energy needs, with electricity consumption from AI data centers surging 50% last year, according to the International Energy Agency. The United Nations University Institute for Water, Environment and Health said that by 2030, AI data centers would use nearly three times the electricity of Nigeria, Pakistan and Bangladesh combined, countries home to some 650 million people.
Tech companies are becoming major energy buyers – and backing renewable and fossil fuel projects. In a speech recently, Simon Stiell, the United Nations climate chief, called the AI industry “energy-guzzling” and said it needs to “respect science and start aligning with global climate efforts – urgently”.
The Turkish environment minister, Murat Kurum, who will preside over this year’s international climate conference in November, said the agenda will include AI’s “growing energy consumption”. He added that “a warming world, a global energy system in disarray and growing fears of the rise of AI all demand action”.
Six years ago, Google had put forth a plan to operate on carbon-free energy by 2030. Google’s CEO, Sundar Pichai, called it a “sustainability moonshot” and referenced the fact that several years earlier, his hometown in India, Chennai, had been submerged by floodwaters, which made “the impacts of climate change feel much closer to home.”
But the AI race has changed the priorities.
Google’s emissions continue to rise. Same for Microsoft, which said it would be carbon-negative by 2030.
Bill Gates, the billionaire co-founder of Microsoft, last year published a 5,000-word essay arguing that climate change would “not lead to humanity’s demise.” His conclusion, made after years of funding climate causes, reflected a much-contested conclusion that even as the planet warms, people would be able to “live and thrive in most places on Earth”.
Then, last month, Gates published another lengthy essay. It was about the risks of AI.
Shared trait: Unpredictability
But many scientists say it’s a mistake to assume that climate change will proceed predictably or in a way that countries can manage. Some say that over the past decade, the planet’s systems have proved more sensitive to incremental warming than anticipated. And plenty more warming is still to come.
The world is losing the beauty of coral reefs, and the fish that call them home, and the economies based on those fish. Plants and trees are withering under tougher conditions, potentially cutting back one of the natural ways that the Earth absorbs carbon dioxide. Melting Arctic sea ice is opening up tracts of dark, heat-absorbing water and further juicing temperature gains. And in the future, changing ocean conditions could weaken or shut down little-considered circulation systems that are vital to maintaining living conditions in some parts of the world.
The risks of feedback loops, and the complexity of the Earth’s system, is precisely why it’s dangerous to make assumptions about how the world will look and how livable it will be, said Katherine Richardson, a University of Copenhagen professor who has studied planetary boundaries.
She said that almost all of human progress – through the Stone Age, the Bronze Age, the Renaissance – has occurred with temperatures remaining in a narrow bound. “We do not know for certain that we can thrive” in a world that is 2°C or 3°C warmer, she said. In fact, she added, we don’t yet know what such a world might look like.
To describe the uncertainties of climate change, she used some of the same language used in warnings about AI.
“We can lose control of this process,” she said. – ©2026 The New York Times Company
This article originally appeared in The New York Times.
