0
0
Artificial Intelligence (AI) has very recently emerged as a powerful tool with the potential to address global challenges, including anthropogenically induced climate change and environmental sustainability. However, its rapid adoption across global industries, education, and recreation has raised concerns about its uneven distribution of environmental impacts. This evaluation of impacts is particularly vital in our changing climate; assessing the damages we are bringing in the development and expansion of AI software now could save a climate crisis of the future—or exacerbate it.
AI as an Accomplice…
Rapid processing 24/7, efficient data analysis, and unmatched information storage—AI offers many significant opportunities to combat climate change and promote sustainability:
Firstly, improved understanding. AI can enhance our comprehension of climate change through advanced data analysis and modelling, crunching millions (or billions) of datasets simultaneously to produce outputs almost instantly. This output can then be used by scientists to adapt their sustainability practices in real-time, avoiding a data time lag in the process of making informed decisions.
Additionally, AI boasts great strengths across a plethora of sector-specific applications. In plant and animal healthcare, AI can be involved in early disease detection and epidemic identification. In agriculture, AI can be used to produce optimised weather forecasting using machine learning software, demonstrated with Google’s “DeepMind” AI platform “GraphCast,” which can predict future weather worldwide more precisely than conventional methods and in less than 60 seconds. This is facilitated by AI software’s ability to run 1,000–10,000 times faster than conventional methods.
Furthermore, AI can provide advice on planting and harvesting when programmed with the farm’s specific resources and inventories, learning from the previous year’s data on production yields, profits, and environmental conditions.
And finally (but not exhaustively), AI has multiple benefits in optimising efficiency within industrial supply chains. Increased efficiency in storage techniques, manufacturing, and distribution systems is all possible and easily accessible to businesses large and small. AI can be accessed on a wide range of budgets, proven by the global adoption of “ChatGTP” in 2024, a free AI software available on any search engine.
AI as an Antagonist…
Despite its benefits, AI poses several environmental challenges. The carbon footprint of training large AI models produces significant emissions. For example, training a single AI model can emit as much CO₂ as 300 round-trip flights between New York and San Francisco.
Energy consumption is another key concern. The global AI energy demand is projected to increase tenfold by 2026, potentially exceeding the annual electricity consumption of small countries such as Belgium. On a more personal level, the International Energy Agency has revealed that a single request made through “ChatGPT” consumes ten times the electricity of a Google Search.
In addition, AI operations exploit responsible levels of water usage. AI systems, particularly data centres, consume huge amounts of water for cooling, which can directly contribute to water scarcity in drought-prone regions. As reported by UNEP, AI-related infrastructure may soon consume six times more water than Denmark, a country of 6 million.
And finally (but also not exhaustively), electronic waste and resource demand are serious considerations in the responsible expansion of AI-enhanced infrastructure. The proliferation of AI hardware contributes to the growing problem of e-waste, which contains hazardous materials that can contaminate soil and water supplies. Additionally, AI relies heavily on critical minerals and rare elements, which are often mined unsustainably.
A Balancing Act: Towards Sustainable AI
The key to sustainable and responsible AI use in industries will be finding methods to harness its potential while mitigating its environmental impact. To develop this, several initiatives are underway, alongside non-binding recommendations for socially and environmentally ethical AI usage, adopted by 190 countries.
Efficient infrastructure can be achieved—advancements in data centre power and cooling can significantly reduce the energy overheads of an industry using AI-enhanced software. In 2022, Google successfully operated its data centre in Finland using 97% carbon-free energy, as announced in their environmental report. By nature, renewable energy is restricted to weather-optimal conditions, proving occasionally unreliable outside of sunny or windy weather—not suitable for the 24/7 high-energy demands of AI infrastructure.
One way to address this is geographical load balancing—a technique that aligns energy demand with real-time grid conditions across globally distributed data centres. This is facilitated by software that considers and evaluates the real-time local factors of each centre, such as the proportion of fossil-based energy sources and water efficiency. Alternating AI traffic across a global network of data centres offers substantial opportunities to tackle AI’s environmental inequality.
Data centre operators are also controversially pursuing “net zero” emissions through carbon offsetting schemes. The carbon offset credits are derived from environmentally beneficial practices, such as solar farms and renewable energy production. While this doesn’t actively reduce the carbon footprints of polluting businesses, it can support those that do.
AI presents both opportunities and challenges for environmental sustainability. While it offers itself as a powerful tool to combat climate change and optimise resource use, its own environmental footprint cannot—and must not—be ignored. The key to responsible development and facilitation of AI technologies will be in the prioritisation of energy and water efficiency to preserve environmental equity. AI development is dynamic; as it continues to evolve, striking the balance between responsible and efficient application will be crucial in ensuring that it remains an accomplice, not an antagonist, to our planet.
Edited by: Veronika Parfjonova
Image: Brain control with artificial intelligence by General, 2018 // Pixabay Content License
Average Rating