With the climate crisis escalating faster than most people expected, one big questions to consider is whether advances in artificial intelligence will unlock new solutions to the problem. For all of the technology's potential, much will depend on how we, the humans, put it to use.
PS Quarterly regularly features predictions by leading thinkers and uniquely positioned commentators on a topic of global concern. Following the public release of powerful new generative AI models late last year, there has been a tidal wave of prognostication about the technology’s potential to transform entire economic sectors, scientific research, education, public-policy planning, and much else. With this broader debate in mind, we asked contributors to respond to the following prompt:
AI will be a game-changer for reaching the Paris climate agreement’s global targets. Agree or disagree?
If we understand AI as a human amplifier, we should not expect it magically to fix climate change. While it could significantly improve climate modeling and help find novel solutions, it also can be expected to deliver much more of the current status quo. What might that look like? One hint comes from journalist Christopher Leonard’s 2019 book Kochland, which describes how Koch Industries profited by delaying the decommissioning of old power plants through a combination of lobbying and climate-change denial.
AI isn’t a policy panacea. It cannot single-handedly restructure institutions that allow people to profit from the planet’s peril. If it functions as an amplifier, it could enable even more ingenious and profitable ways to subvert policies designed to achieve our climate targets.
As a general matter, we should avoid the magical thinking promulgated by tech billionaires. Listen to them, and you might think that we don’t need to worry too much about climate change, because Elon Musk will build a new civilization on Mars by 2050. Obviously, technology will provide new solutions to our problems. But it also helped us create those problems in the first place.
I hope we do capitalize on the disruptive effects of AI. Rather than augmenting business as usual, perhaps it could instead prompt us to rethink our priorities, and to change our policies and institutions accordingly.
Secure your copy of PS Quarterly: The Year Ahead 2025
Our annual flagship magazine, PS Quarterly: The Year Ahead 2025, has arrived. To gain digital access to all of the magazine’s content, and receive your print copy, subscribe to PS Digital Plus now.
AI will surely help us achieve the Paris climate targets. After all, it already has many proven applications for monitoring emissions and energy usage. Researchers at the University of Maryland have been relying on machine-learning techniques since the early 2000s to detect deforestation in the Amazon rainforest from satellite data. And in 2016, Google DeepMind developed a machine-learning system that reduced the energy used in data-center cooling by 40%.
But AI use remains low in many contexts. Systems to monitor many sources of emissions, such as agricultural methane, are only now being developed, and many businesses worldwide lack the data, money, and skilled personnel to adopt AI. That is where new large language models come in, because they hold the promise of enabling far more people to incorporate machine learning and other types of AI into their current work. In fact, clean-energy professionals who lack coding expertise are already using tools like ChatGPT to develop systems for optimizing utility-scale battery storage deployment.
These uses could significantly accelerate clean-tech development. Last year, Google DeepMind reported that it had trained a first-of-its-kind reinforcement-learning algorithm to control plasma in nuclear fusion reactors. Just imagine the advances that are possible by 2050 as millions of scientists, engineers, entrepreneurs, and others gain access to these tools, whose own capabilities will only continue to grow.
AI is undeniably the game-changer needed to help us achieve the Paris agreement’s targets. Its potential lies in its ability to crunch vast amounts of data, discern patterns, and predict behavior with unmatched accuracy. By harnessing this predictive power, we can unlock deeper insights into the intricate links between climate change and the biodiversity crisis. These two existential issues fuel each other and cannot be solved in isolation, yet the existing international agreements to address them operate independently.
We can deploy AI to develop an integrated approach that unites international accords on climate change, biodiversity, and a range of other issues. Armed with AI-driven analyses of thousands of treaties, we can avoid duplication, resolve disputes, and ensure comprehensive coverage of critical issues. Moreover, inclusive negotiations are crucial for meaningful progress on climate change, and AI could be a great equalizer, giving developing countries access to cutting-edge data analysis. Imagine the results if smaller countries wielded the same analytical power as their larger counterparts at the negotiating table.
That said, we must remember that AI is not a substitute for human ingenuity and moral judgment. Safeguarding against biases and unethical outcomes needs to be a top priority. As the climate crisis escalates, we should seize on every opportunity that presents itself. AI introduces unparalleled possibilities to shape the agenda for UN Climate Week, COP28, and beyond. For the sake of our future on this planet, we must embrace it.
AI will have a detrimental effect on our climate targets because it is already yet another engine of ecocide. As the next extension of an already highly destructive capitalist technological system, AI will soak up fossil fuels (through energy use), rare metals, land, and even water (to cool servers), while externalizing pollution. Moreover, AI systems will further consolidate power and wealth in the hands of a few corporations, which are amassing data for use in surveillance and the exploitation of human labor – all for short-term profit.
To counter AI-ecocide, we must mount successful interventions at all levels – from international rule-making down to our own personal education. Most people do not associate the digital realm with encroachment on planetary boundaries. That needs to change. To reduce energy and resource consumption to sustainable levels, we must resist the runaway production and operation of more digital devices, networking infrastructure, and data centers. Extractivism and growth-oriented business models must be countered with economic policies that justly redistribute the benefits of technology. Technical innovation must focus more on sustaining critical systems and institutions, and on repairing harms.
If AI development continues down the track of anthropocentric, hierarchical, individualistic applications, it will not end well. But if we acknowledge other forms of “real intelligence” in nature – swarms, reefs, hives, forests, octopuses – that could instill us with the humility we need to change course.
As the world grapples with climate change and its effects, we confront the need to make decisions about mitigation and adaptation measures with uncertain and incomplete information. AI can help us understand what’s happening now and where we’re headed, so that we know where to target our efforts and investment.
In the Johns Hopkins University Applied Physics Laboratory’s (APL) work as part of the Climate TRACE coalition, we’re using satellite imagery, road networks, and machine learning to estimate the amount of greenhouse-gas (GHG) emissions coming from road transportation around the world. Last year, we released emissions estimates for 500 major urban areas, many of which previously did not have this data. This fall, we will expand to thousands of cities – an increase in coverage made possible by AI. Such information shows where GHGs are being released so that policymakers can set more targeted goals and track progress.
AI can also be used to accelerate our broader understanding of climate change, ultimately creating an impetus for changes that will help us avoid or lessen its impact. A team of APL researchers is using generative AI to discover drivers for climate tipping points, starting with the potential collapse of the Atlantic Meridional Overturning Circulation. Discovering these underlying factors yields better insights into what temperature goals we must meet to prevent reaching tipping points – not to mention providing additional motivation for pursuing climate action.
When we incorporate AI into the monitoring and modeling of our Earth systems, we enhance our understanding and are better positioned to identify solutions for achieving a healthier planet.
To have unlimited access to our content including in-depth commentaries, book reviews, exclusive interviews, PS OnPoint and PS The Big Picture, please subscribe
PS Quarterly regularly features predictions by leading thinkers and uniquely positioned commentators on a topic of global concern. Following the public release of powerful new generative AI models late last year, there has been a tidal wave of prognostication about the technology’s potential to transform entire economic sectors, scientific research, education, public-policy planning, and much else. With this broader debate in mind, we asked contributors to respond to the following prompt:
AI will be a game-changer for reaching the Paris climate agreement’s global targets. Agree or disagree?
Nicholas Agar
If we understand AI as a human amplifier, we should not expect it magically to fix climate change. While it could significantly improve climate modeling and help find novel solutions, it also can be expected to deliver much more of the current status quo. What might that look like? One hint comes from journalist Christopher Leonard’s 2019 book Kochland, which describes how Koch Industries profited by delaying the decommissioning of old power plants through a combination of lobbying and climate-change denial.
AI isn’t a policy panacea. It cannot single-handedly restructure institutions that allow people to profit from the planet’s peril. If it functions as an amplifier, it could enable even more ingenious and profitable ways to subvert policies designed to achieve our climate targets.
As a general matter, we should avoid the magical thinking promulgated by tech billionaires. Listen to them, and you might think that we don’t need to worry too much about climate change, because Elon Musk will build a new civilization on Mars by 2050. Obviously, technology will provide new solutions to our problems. But it also helped us create those problems in the first place.
I hope we do capitalize on the disruptive effects of AI. Rather than augmenting business as usual, perhaps it could instead prompt us to rethink our priorities, and to change our policies and institutions accordingly.
Secure your copy of PS Quarterly: The Year Ahead 2025
Our annual flagship magazine, PS Quarterly: The Year Ahead 2025, has arrived. To gain digital access to all of the magazine’s content, and receive your print copy, subscribe to PS Digital Plus now.
Subscribe Now
Dan Blaustein-Rejto
AI will surely help us achieve the Paris climate targets. After all, it already has many proven applications for monitoring emissions and energy usage. Researchers at the University of Maryland have been relying on machine-learning techniques since the early 2000s to detect deforestation in the Amazon rainforest from satellite data. And in 2016, Google DeepMind developed a machine-learning system that reduced the energy used in data-center cooling by 40%.
But AI use remains low in many contexts. Systems to monitor many sources of emissions, such as agricultural methane, are only now being developed, and many businesses worldwide lack the data, money, and skilled personnel to adopt AI. That is where new large language models come in, because they hold the promise of enabling far more people to incorporate machine learning and other types of AI into their current work. In fact, clean-energy professionals who lack coding expertise are already using tools like ChatGPT to develop systems for optimizing utility-scale battery storage deployment.
These uses could significantly accelerate clean-tech development. Last year, Google DeepMind reported that it had trained a first-of-its-kind reinforcement-learning algorithm to control plasma in nuclear fusion reactors. Just imagine the advances that are possible by 2050 as millions of scientists, engineers, entrepreneurs, and others gain access to these tools, whose own capabilities will only continue to grow.
Maxwell Gomera
AI is undeniably the game-changer needed to help us achieve the Paris agreement’s targets. Its potential lies in its ability to crunch vast amounts of data, discern patterns, and predict behavior with unmatched accuracy. By harnessing this predictive power, we can unlock deeper insights into the intricate links between climate change and the biodiversity crisis. These two existential issues fuel each other and cannot be solved in isolation, yet the existing international agreements to address them operate independently.
We can deploy AI to develop an integrated approach that unites international accords on climate change, biodiversity, and a range of other issues. Armed with AI-driven analyses of thousands of treaties, we can avoid duplication, resolve disputes, and ensure comprehensive coverage of critical issues. Moreover, inclusive negotiations are crucial for meaningful progress on climate change, and AI could be a great equalizer, giving developing countries access to cutting-edge data analysis. Imagine the results if smaller countries wielded the same analytical power as their larger counterparts at the negotiating table.
That said, we must remember that AI is not a substitute for human ingenuity and moral judgment. Safeguarding against biases and unethical outcomes needs to be a top priority. As the climate crisis escalates, we should seize on every opportunity that presents itself. AI introduces unparalleled possibilities to shape the agenda for UN Climate Week, COP28, and beyond. For the sake of our future on this planet, we must embrace it.
Vesna Manojlovic
AI will have a detrimental effect on our climate targets because it is already yet another engine of ecocide. As the next extension of an already highly destructive capitalist technological system, AI will soak up fossil fuels (through energy use), rare metals, land, and even water (to cool servers), while externalizing pollution. Moreover, AI systems will further consolidate power and wealth in the hands of a few corporations, which are amassing data for use in surveillance and the exploitation of human labor – all for short-term profit.
To counter AI-ecocide, we must mount successful interventions at all levels – from international rule-making down to our own personal education. Most people do not associate the digital realm with encroachment on planetary boundaries. That needs to change. To reduce energy and resource consumption to sustainable levels, we must resist the runaway production and operation of more digital devices, networking infrastructure, and data centers. Extractivism and growth-oriented business models must be countered with economic policies that justly redistribute the benefits of technology. Technical innovation must focus more on sustaining critical systems and institutions, and on repairing harms.
If AI development continues down the track of anthropocentric, hierarchical, individualistic applications, it will not end well. But if we acknowledge other forms of “real intelligence” in nature – swarms, reefs, hives, forests, octopuses – that could instill us with the humility we need to change course.
Elizabeth Reilly
As the world grapples with climate change and its effects, we confront the need to make decisions about mitigation and adaptation measures with uncertain and incomplete information. AI can help us understand what’s happening now and where we’re headed, so that we know where to target our efforts and investment.
In the Johns Hopkins University Applied Physics Laboratory’s (APL) work as part of the Climate TRACE coalition, we’re using satellite imagery, road networks, and machine learning to estimate the amount of greenhouse-gas (GHG) emissions coming from road transportation around the world. Last year, we released emissions estimates for 500 major urban areas, many of which previously did not have this data. This fall, we will expand to thousands of cities – an increase in coverage made possible by AI. Such information shows where GHGs are being released so that policymakers can set more targeted goals and track progress.
AI can also be used to accelerate our broader understanding of climate change, ultimately creating an impetus for changes that will help us avoid or lessen its impact. A team of APL researchers is using generative AI to discover drivers for climate tipping points, starting with the potential collapse of the Atlantic Meridional Overturning Circulation. Discovering these underlying factors yields better insights into what temperature goals we must meet to prevent reaching tipping points – not to mention providing additional motivation for pursuing climate action.
When we incorporate AI into the monitoring and modeling of our Earth systems, we enhance our understanding and are better positioned to identify solutions for achieving a healthier planet.