AI for Earth: New weapon in climate change mitigation arsenal

By Seda Sevencan

ISTANBUL (AA) — Since artificial intelligence (AI) erupted onto the scene, experts and scientists have been making inroads into using its potential to help humanity avoid climate calamity.

Billions of people around the world face growing risks as the planet continues to approach 1.5 C of global warming, already manifesting in devastating disasters from flooding to droughts and heatwaves.

Today, scientists and experts from diverse fields are devoting vast resources, expertise, and ingenuity in the quest to address this planet-wide threat, with AI, unsurprisingly, emerging as a promising avenue towards expediting this effort.

According to a 2023 report by the Boston Consulting Group (BCG) and Google, AI can potentially mitigate between 5-10% of global greenhouse emissions by 2030.

One of the ways AI can help make this happen is by sifting through massive expanses of digital data and finding patterns and insights that may prove crucial to assessing the challenge and potential solutions, according to Alp Kucukelbir, an adjunct professor of computer science at Columbia University.

Besides being the chief scientist at Fero Labs, a company that offers AI-driven manufacturing process optimization software, Kucukelbir is also a member of Climate Change AI, a member of an international network of academics, industry members, and policymakers focused on the role machine learning can play in tackling climate change.

“Machine learning and AI technology is very good at taking very large amounts of data and finding the signal, the important parts, the salient parts from very large data,” Kucukelbir tells Anadolu in an interview on how AI can help, as well as hurt, climate action and policy.

Besides detecting carbon and methane leakage, crunching mountains of data and scientific papers through AI empowers policymakers to make informed choices and implement effective climate strategies, according to Kucukelbir, who said:

“AI can help understand and access a much larger amount of information that helps us make better decisions about what to do about the climate.”

From reducing energy consumption and waste to increasing the circularity of our economy and adapting to new realities, AI can help cut down on our impact on the environment.

“It could be the heating and cooling of buildings. It could be the optimizing the route of maritime shipping so that you use less gas to make the same route. It could be the optimization of traffic lights to minimize the number of times cars stop and go up. So, all of these are cases where you're reducing the amount of energy, but you're achieving the same outcome,” explained Kucukelbir.

“The food system is a very good example of this. We use enormous amounts of fertilizer, herbicide, pesticide as industrial scale to make the food that we use. Using AI, you can really tailor exactly how much is needed so that you don't use excess amounts of these kinds of things.”

- Invaluable tool, no silver bullet

Yet, as with any powerful tool, AI comes with its challenges and considerations.

According to Kucukelbir, “the first thing to make sure that's very clear is that what AI cannot do.”

While AI has the potential to become an invaluable tool for climate action, it is important not to lose sight of the fact that it will still be humans who will have to actually take that action.

“AI, machine learning, software cannot physically move molecules and physically change things,” he said.

“When we think about climate change mitigation, we need to remove the carbon in the air and we need to put less carbon in the air. This is the movement of carbon. AI is not directly doing these things.”

Both AI’s limitations and its strengths mean point to the importance of gathering data and ensuring that it is available to those who need to use it.

“AI is a data driven software technology. If we don't have data in a particular application area, we can't build good promising AI solutions,” says Kucukelbir.

He and other experts have emphasized the need for governments to incentivize data gathering, standardization, and sharing across universities, companies, governments, and others to “make sure that we have a common area, a foundation for data.”

“But the biggest challenge in practice is the people and the talent challenge,” he underlined.

For Kucukelbir, collaboration between experts in various domains and AI practitioners is key to unlocking AI's full potential in the fight against climate change.

“There are people who have dedicated their entire careers to specialty knowledge about energy systems, about making cement, about growing wheat, things like that. These are not the same people who have computer science, machine learning, AI expertise.

“And on the other side, machine learning, AI experts, they don't understand the actual domain application, they are not experts in those things,” he said.

“The biggest barrier there is to make sure that these two sets of people have effective ways of talking to each other, collaborating with each other. And my view there is that the tools that are being developed today are enabling that interaction going forward.”

- Carbon footprint and expert collaboration

Ensuring collaboration will be crucial for the future of AI’s role addressing climate change, especially in overcoming ways that this technology could potentially impede efforts to stop global warming.

Besides the carbon footprint of AI itself, Kucukelbir acknowledges the dual nature of AI, capable of either amplifying harmful practices or driving sustainable solutions. “Just like any other technology, AI can be used to make the matter worse,” he said.

“What is the protection against that? It is the same protection that we have against all technology is to have a system that penalizes and disincentivizes such actions.”

On the emissions that AI technology causes, Kucukelbir said this is also a critical concern as it rapidly becomes more widespread.

One study that he was involved in showed that currently, greenhouse emissions due to AI are significantly less than 1% of the global total, though it remains uncertain what this figure will be in the future.

- AI-enmeshed ‘operating system of the world’

Peering into the future, Kucukelbir paints a compelling picture of AI's role in shaping global sustainability efforts. Envisioning a world where it is seamlessly integrated into every facet of our lives. He foresees a future where technology guides us towards more efficient resource usage and sustainable practices.

“In the next five to 10 years, I see machine learning and AI as being quite literally part of the so-called operating system of the world. Every single sector of the economy will have AI and machine learning in it that is guiding us humans to make better decisions,” he said.

“They will be steering us towards minimizing the amount of waste, minimizing the amount of energy that we're using, but also making us more effective and efficient. We're going to be able to make more with less resources, make the same amount of things, faster.”​​​​​​​