AI in Climate Action: Between Harnessing Potential and Navigating Challenges

Transforming Agriculture and Energy

AI has already begun to reshape key sectors such as agriculture and renewable energy, where significant progress can be made toward achieving sustainability goals:

• Sustainable Agriculture: AI systems can enhance the efficiency of agricultural production by optimising resource use. Advanced predictive models analyse soil data, weather patterns, and crop behaviour to determine the precise amount of water, fertiliser, and pesticides needed. This reduces waste, cuts costs, and minimises the environmental impact of agriculture, making it better suited to a changing climate. According to the study Tackling Climate Change with Machine Learning, machine learning models can also aid in precision agriculture by analysing remote sensing data to monitor crop health and identify pests or diseases early, which further improves yields and resource efficiency.
• Renewable Energy Optimisation: AI’s role in energy systems is even more promising. By analysing complex weather data, AI can improve the efficiency and integration of solar and wind energy into the grid. Predictive models enhance the accuracy of energy production forecasts, enabling grid operators to balance supply and demand more reliably and reduce dependency on fossil fuels. As highlighted in the above-mentioned study, AI can also optimise energy storage systems, enabling better use of renewable energy by predicting when and where energy will be needed most, minimising curtailment of excess power.

AI’s contributions to these sectors demonstrate its ability to make systems more efficient, scalable, and adaptable, potentially offering a powerful boost in the fight against climate change.

A ChatGPT query can use up to 10 times more electricity than a traditional Google search, depending on the complexity and output format.

European Parliamentary Research Service (EPRS)

AI Can Pose Significant Environmental and Social Risks

However, AI is not a one-size-fits-all solution. The development and deployment of AI technologies come with environmental costs and societal risks that must not be ignored.

• High Energy Consumption: Training advanced AI models requires significant computational power, which, in turn, consumes vast amounts of energy. For instance, Google reported a 27% year-on-year increase in data center electricity consumption, driven by business growth and rising product adoption, including AI. This creates a paradox where the technologies meant to help mitigate climate change risk contribute to it by driving up emissions. Research shows that training a single large AI model can emit as much carbon as five cars over their entire lifetimes. 
• Resource Inequities: The benefits of AI are not evenly distributed, particularly in the Global South, where access to AI infrastructure and data is limited. Projects like the GIZ initiative “Fair Forward – Artificial Intelligence for All”  aim to address this disparity, promoting fair and inclusive access to AI technologies. Nonetheless, more needs to be done to ensure that AI contributes to equitable and sustainable development worldwide.

“Machine learning or AI is one tiny piece of the puzzle. You already have a climate solution that is much broader, and you have analytical approaches to it. Then machine learning gives that new ability to it, which makes it so exciting. But it’s not that AI just comes in and fixes things.”

Lynn Kaack, leading expert in the field of AI and Climate Change and co-author of the summarised study, in her keynote presentation at the EUKI Networking Conference 2024

The Way Forward: Ensuring AI Becomes a Climate Ally

The question remains: How can AI become a climate ally? To ensure AI fulfils its promise three key areas must be addressed, according to the authors of Tackling Climate Change with Machine Learning:

1. Green AI Development: The environmental impact of AI systems must be minimised. Researchers and developers must prioritise energy-efficient algorithms, explore hardware innovations, and rely on renewable energy to power data centres. For example, developing low-energy AI models and improving training efficiency can significantly reduce AI’s carbon footprint.
2. Global Accessibility: Governments and organisations should invest in equitable AI access, fostering collaboration between developed and developing regions. Initiatives such as knowledge-sharing networks and open-source AI tools can empower countries to use AI in their climate strategies. ML techniques for low-data settings, such as transfer learning, are particularly relevant for regions lacking large datasets or technical infrastructure.
3. Policy and Regulation: Policymakers must ensure that AI solutions align with climate goals. Strong regulations on energy consumption in AI development, combined with incentives for low-carbon AI projects, will drive innovation while reducing environmental harm. For instance, policies that prioritise AI applications for climate mitigation and adaptation over other less impactful uses can help maximise AI’s societal benefits.

Renewables and natural gas are the main energy sources used to power data centres globally, although nuclear (and, in the future, small modular reactors) is also on the rise.

European Parliamentary Research Service (EPRS)

Conclusion: A Critical Opportunity for Climate Action

AI is a transformative force that holds the potential to become a critical tool in addressing the climate crisis. It can optimise agriculture, accelerate renewable energy adoption, and offer solutions for mitigation and adaptation. Yet, it is not without risks. To ensure AI is a positive force for climate action, we must address its energy demands, foster equitable access, and integrate strong policies that align AI development with climate goals.

In the EU, data centres account for approximately 3% of total electricity consumption, though this varies by country and surpasses 20% in Ireland.

European Parliamentary Research Service (EPRS)

As we look ahead, one thing is clear: AI’s role in the climate fight must be approached thoughtfully. If we act responsibly, AI can indeed become a cornerstone of sustainable innovation – a powerful ally in safeguarding the future of our planet.

Source: Rolnick, D., Donti, P. L., Kaack, L. H., et al. Tackling Climate Change with Machine Learning. ACM Computing Surveys, Vol. 55, No. 2, 2022.

Website that accompanies this article, where the authors offer additional resources, as well as opportunities for knowledge-sharing and networking: www.climatechange.ai