AI is anticipating a temperature rise as high as 3 degrees by the end of the century. Could it be right?
A United Nations report released late November predicted a 2.5-2.9 degree rise in global temperature this century compared to pre-industrial levels. Various researchers at Stanford and the UK Met Office, as well as startups like ClimateAI, have simulated various scenarios in which global temperatures look set to soar.
However, as AI emerges as a decent forecaster and potential troubleshooter, a broader paradox emerges – it might be part of the problem it’s trying to solve.
AI predicts a much faster warming
A recent study led by Elizabeth Barnes, Noah Diffenbaugh, and Sonia Seneviratne used AI-powered transfer learning to refine regional climate predictions. This advanced technique, which combines insights from 10 global climate models with real-world observations, revealed that 34 regions will likely exceed the critical 1.5°C warming threshold by 2040.
By 2060, 26 of those regions are projected to surpass 3°C of warming—much faster than earlier predictions.
Vulnerable areas like South Asia, the Mediterranean, Central Europe, and sub-Saharan Africa are expected to face accelerated warming, heightening risks to ecosystems, water resources, and communities.
The researchers emphasized the importance of focusing on regional impacts, where climate changes are often more extreme and unpredictable due to complex local processes.
The study highlights how AI can significantly improve regional climate forecasting, providing policymakers with critical insights for better adaptation strategies.
These findings reinforce the urgency of addressing climate risks not just globally but also locally, with targeted measures to protect the most affected regions.
The promise and paradox of AI in climate forecasting
One particularly important element is how AI can accurately forecast future implications for climate change. The CMIP6 model – which simulates Earth's climate to predict future warming scenarios – applies more accurate physics, chemistry, and biology to assess climate sensitivity.
However, AI doesn't get a free pass when it comes to energy consumption. According to a recent podcast by Climactic co-founder Raj Kapor, data centers are projected to account for around 6% of the US’s total energy consumption over the next couple of years. Climactic is a venture investment firm for climate tech.
Also, and somewhat paradoxically, AI is being used in the US Department of Energy’s Office of Fossil Energy and Carbon Management to help streamline the industry, which directly undermines the idealistic notion that AI is totally geared towards going green.
It seems that AI can help, with precision irrigation systems in agriculture just one of many possible examples mentioned.
In terms of investment, AI can also act as an accountant, calculating the most cost-effective solar panels to invest in, as businesses are often reluctant to make these decisions without hard data.
As businesses emerge as key arbiters of change, Kapor also mentioned that AI is capable of being the ultimate salesperson, helping companies streamline their carbon footprint.
He shared an example of AI analyzing objections from individuals and businesses, then building a “loop of engagement” using images, presentations, and videos to sell a green path forward.
However, there are still limitations. Kapor mentioned that climate tech currently lacks precise net reduction calculations to convince investors, but efforts to improve this are underway, and it could prove vital to the climate conversation.
Safeguarding AI infrastructure and combating greenwashing
There are particular vulnerabilities in the data centers that power AI. As the weather becomes more extreme, these facilities face increasing risks from floods, wildfires, and heatwaves, especially as many are in the desert.
Still, AI could play a crucial role in safeguarding these sites by predicting risks and optimizing infrastructure.
We reached out to Kelwin Fernandes, CEO of NILG.AI, a business optimization startup in Portugal. He shared the following:
“I don't believe a single action (AI or politics) will make a major impact on the environment on a global scale. We must aim for small local optimizations in our specific industries and areas. For instance, we recently worked on an AI project to determine the best location of a recycling bin, expanding the network coverage while facilitating operations.”
So leaving the burden of solution entirely on AI is irresponsible and doesn’t give us the license to slack off. Leveraging AI as a monitoring tool for major corporations like Amazon or Google seems like a crucial step on the long path.
AI can also combat greenwashing – when companies falsely portray themselves as environmentally friendly to mislead consumers – by analyzing corporate claims about sustainability. It could sift through company reports, production data, and supply chain activities to expose discrepancies or deceptive practices.
Your email address will not be published. Required fields are markedmarked