The Paradox
AI’s biggest trick is convincing us it lives in the cloud. It doesn’t. It lives in vast, humming sheds of servers that run so hot they need their own rivers. Every time a chatbot writes a haiku, somewhere a turbine is spinning faster.
And that’s the problem: we’ve created the most powerful cognitive tool in human history, and it lives on a diet of fossil fuels and freshwater. Estimates suggest AI’s total energy demand could increase tenfold within five years. Not ten percent. Tenfold.
That sounds bleak, but there is another side to the story. The same technology driving this consumption could be the very thing that saves us in the battle against climate change. Given this, can we afford to be without it?
Absolutely not.
But we must act quickly. Rolling out AI-powered climate change countermeasures at a global scale is a vital and urgent next step. In a battle we as a species are gradually losing with each month that passes, only immediate action can help us overcome the challenges that lie ahead.
The Energy Appetite
It is estimated that over a billion people use generative AI tools every day. With each interaction consuming around 0.34 watt-hours per prompt, we’re talking about 310 gigawatt-hours every year. That is equivalent to the annual electricity use of over 3 million people in a low-income African country. The training of GPT-4 consumed over 50 gigawatt-hours of electricity, enough to power San Francisco for three consecutive days.
Most of the energy is spent powering the vast mega-banks of GPUs that power AI systems, but there is also the additional energy drain of cooling these systems, which generate vast amounts of heat.
Since the inception of what we now know as AI, the industry has been focused heavily on making it more efficient. Be it reducing model size, improving chip performance, or using renewables to power data centers, all attempts can and should be implemented to help reduce energy consumption.
However, a more sustainable AI industry isn’t solely about reducing kilowatt-hours; it’s about producing actionable ideas that can help restore planetary health.
Pulling the plug on AI won’t save the planet; it’s the only tool powerful enough to save it. The challenge is to direct the extreme levels of computational power toward outcomes that justify the cost and to deliver what we need before it’s too late.
The Opportunity
The ‘I’ in ‘AI’ is irrefutable. The ability for AI models to see patterns and optimize makes them significantly superior to human minds when it comes to unlocking efficiencies.
The rate at which intelligence is progressing is also notable. In 2023, researchers introduced three benchmarks to measure improved AI performance (MMMU, GPQA, and SWE-bench). In the following year, performance increased 18.8, 48.9, and 67.3 percentage points, respectively.
AI is already being successfully implemented to model climate patterns and predict weather far better than we humans are capable of. Be it Spherical DYffusion, which can predict climate patterns over 100 years 25 times faster than the previous norm, or DLESyM, which can simulate 1,000 years of the Earth’s present climate conditions in just one day. These aren’t theoretical or potential future uses of AI; they’re happening as we speak.
The overall view of AI and climate change shouldn’t be one of polar extremes, as we so often see elsewhere in society today. The answer isn’t banning AI to save energy, and it’s not endlessly using AI for random, unconsidered reasons. There is a middle ground where the AI potential and sustainability converge for the better.
As climate change continues to evolve and cause damage, the most valuable intelligence will be the kind that drives real-world change. Given this is so, AI-powered regeneration won’t be a luxury; it’ll be a necessity.
Given the existing proof points, now is the time to trust the genius we’ve created and roll out AI-based solutions as quickly as possible while ensuring they are effective and safe, to combat the almost daily headlines about irreversible harm to the environment.
The Future
Regardless of personal opinions, Sam Altman was right about one thing: there’s never been a better time to start a company. AI has torn down what used to be barriers to entry for entrepreneurs and innovators; what once required an entire workforce to operate can now be done by a single individual focused on precision and speed of output.
This democratization is the climate opportunity. If all emerging AI companies were focused not only on valuation but on carbon removal, the combined global effect would be transformational. Yes, AI has a significant energy bill, but its potential is far more substantial. If every watt spent training a language model is put towards finding solutions to offset climate change, AI’s footprint becomes a down payment on a cleaner future.
AI is arguably the most intelligent thing humanity has ever developed; it only makes sense to put that intelligence towards solving humanity’s biggest problem.
